JP7510913B2 - Highly multiplexed PCR methods and compositions - Google Patents

Description

CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of and priority to U.S. Utility Application No. 13/683,604, filed November 21, 2012, and U.S. Provisional Application No. 61/675,020, filed July 24, 2012. U.S. Utility Application No. 13/683,604 is a continuation-in-part of U.S. Utility Application No. 13/300,235, filed November 18, 2011, which is a continuation-in-part of U.S. Utility Application No. 13/110,685, filed May 18, 2011, and claims the benefit of U.S. Provisional Application No. 61/675,020, filed July 24, 2012. U.S. Utility Application No. 13/110,685 claims the benefit of U.S. Provisional Application No. 61/395,850, filed May 18, 2010; U.S. Provisional Application No. 61/398,159, filed June 21, 2010; U.S. Provisional Application No. 61/462,972, filed February 9, 2011; U.S. Provisional Application No. 61/448,547, filed March 2, 2011; and U.S. Provisional Application No. 61/516,996, filed April 12, 2011. U.S. Utility Application No. 13/300,235 claims the benefit of U.S. Provisional Application No. 61/571,248, filed June 23, 2011. All of these patent applications are incorporated herein by reference in their entirety for all their teachings.

STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT This invention was made with support from the National Institutes of Health under Grant No. 5R44HD60423-3. The United States Government may reserve rights to any patents originating from this application.

FIELD OF THEINVENTION The present invention relates generally to methods and compositions for simultaneously amplifying multiple nucleic acid regions of interest in one reaction volume.

To increase assay throughput and allow for more efficient use of nucleic acid samples, simultaneous amplification of many target nucleic acids in a sample of interest can be performed by mixing many oligonucleotide primers with the sample and then subjecting the sample to polymerase chain reaction (PCR) conditions in a process known in the art as multiplex PCR. The use of multiplex PCR is a very simple experimental procedure and can reduce the time required for nucleic acid analysis and detection. However, when multiple pairs are added to the same PCR reaction, non-target amplification products such as amplified primer dimers may be generated. The likelihood of such product generation increases with increasing number of primers. These non-target amplification products greatly limit the use of the amplification products in subsequent analyses and/or assays. Therefore, improved methods are needed to reduce the formation of non-target amplification products during multiplex PCR.

Improved multiplex PCR methods will be useful for a variety of applications, such as non-invasive prenatal genetic diagnosis (NPD). Specifically, current methods of prenatal diagnosis can alert physicians and parents to abnormalities in the developing fetus. Without prenatal diagnosis, one in 50 infants is born with significant physical or mental handicaps, and as many as one in 30 have some form of congenital malformation. Unfortunately, standard methods involve invasive procedures that are either poorly accurate or carry the risk of miscarriage. Methods based on maternal blood hormone levels or ultrasound measurements are non-invasive but similarly inaccurate. Methods such as amniocentesis, chorionic villus biopsy, and fetal blood sampling are highly accurate but invasive and carry significant risks. Amniocentesis was performed for approximately 3% of all pregnancies in the United States, but its frequency of use has declined over the past 15 years.

Normal humans have two sets of the 23 chromosomes in every healthy diploid cell, one copy from each parent. Aneuploidy, a condition in which a nuclear cell has too many and/or too few chromosomes, is thought to be responsible for the majority of implantation failures, miscarriages, and genetic diseases. Detecting chromosomal abnormalities can identify individuals or embryos with conditions such as Down's syndrome, Klinefelter's syndrome, and Turner's syndrome, among others, in addition to increasing the chances of a successful pregnancy. Testing for chromosomal abnormalities is particularly important since it is estimated that at least 40% of embryos are abnormal between the ages of 35 and 40, and more than half of embryos are abnormal above the age of 40.

It has recently been discovered that cell-free fetal DNA and intact fetal cells can enter the maternal blood circulation. Thus, analysis of this genetic material may enable early NPD. Improved methods would be desirable to increase sensitivity and specificity and reduce the time and cost required for NPD.

In one aspect, the invention features a method of amplifying target loci in a nucleic acid sample. In some embodiments, the method includes (i) contacting the nucleic acid sample with a library of test primers that simultaneously hybridize to at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci to generate a reaction mixture, and (ii) subjecting the reaction mixture to primer extension reaction conditions to generate an amplification product that includes a target amplification product. In some embodiments, the method also includes determining the presence or absence of at least one target amplification product (e.g., at least 50, 60, 70, 80, 90, 95, 96, 97, 98, 99, or 99.5% of the target amplification products). In some embodiments, the method also includes determining the sequence of at least one target amplicon (e.g., at least 50, 60, 70, 80, 90, 95, 96, 97, 98, 99, or 99.5% of the target amplicon).

In various embodiments of any aspect of the invention, at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci are amplified. In some embodiments, at least 50, 60, 70, 80, 90, 95, 96, 97, 98, 99, or 99.5% of the amplification products are target amplification products. In some embodiments, at least 50, 60, 70, 80, 90, 95, 96, 97, 98, 99, or 99.5% of the target loci are amplified. In various embodiments, less than 60, 50, 40, 30, 20, 10, 5, 4, 3, 2, 1, 0.5, 0.25, 0.1, or 0.05% of the amplification products are primer dimers. In some embodiments, the library of test primers includes at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 test primer pairs, each primer pair including a forward test primer and a reverse test primer that hybridize to the same target locus. In some embodiments, the library of test primers includes at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 individual test primers that hybridize to different target loci, and the individual primers are not part of a primer pair.

In various embodiments of any of the aspects of the invention, the concentration of each test primer is less than 100, 75, 50, 25, 10, 5, 2, or 1 nM. In various embodiments, the GC content of the test primers is 30-80%, e.g., 40-70% or 50-60%. In some embodiments, the range of GC content of the test primers (e.g., maximum GC content - minimum GC content, e.g., 80%-60% = 20% range) is less than 30, 20, 10, or 5%. In some embodiments, the melting temperature (T _m ) of the test primers is 40-80°C, e.g., 50-70°C, 55-65°C, or 57-60.5°C. In some embodiments, the range of melting temperatures of the test primers is less than 20, 15, 10, 5, 3, or 1°C. In some embodiments, the length of the test primer is 15-100 nucleotides, e.g., 15-75 nucleotides, 15-40 nucleotides, 17-35 nucleotides, 18-30 nucleotides, 20-65 nucleotides. In some embodiments, the test primer includes a tag that is not target specific, e.g., a tag that forms an internal loop structure. In some embodiments, the tag is between two DNA binding regions. In various embodiments, the test primer includes a 5' region that is specific for the target locus, an internal region that forms a loop structure that is not specific for the target locus, and a 3' region that is specific for the target locus. In various embodiments, the length of the 3' region is at least 7 nucleotides. In some embodiments, the length of the 3' region is 7-20 nucleotides, e.g., 7-15 nucleotides, or 7-10 nucleotides. In various embodiments, the test primer includes a 5' region that is not specific for the target locus (e.g., a tag or universal primer binding site), followed by a region that is specific for the target locus, an internal region that forms a loop structure that is not specific for the target locus, and a 3' region that is specific for the target locus. In some embodiments, the range of lengths of the test primers is less than 50, 40, 30, 20, 10, or 5 nucleotides. In some embodiments, the length of the target amplicons is 50-100 nucleotides, e.g., 60-80 nucleotides, or 60-75 nucleotides. In some embodiments, the range of lengths of the target amplicons is less than 50, 25, 15, 10, or 5 nucleotides.

In various embodiments of any aspect of the invention, the primer extension reaction conditions are polymerase chain reaction conditions (PCR). In various embodiments, the length of the annealing step is greater than 3, 5, 8, 10, or 15 minutes. In various embodiments, the length of the extension step is greater than 3, 5, 8, 10, or 15 minutes.

In various embodiments of any aspect of the invention, test primers are used to simultaneously amplify at least 1,000 different target loci in a sample that includes maternal DNA from the pregnant mother of a fetus and fetal DNA to determine the presence or absence of a fetal chromosomal abnormality. In various embodiments, the method includes ligating a universal primer binding site to DNA molecules in the sample, amplifying the ligated DNA molecules with at least 1,000 specific primers and a universal primer to generate a first set of amplification products, and amplifying the first set of amplification products with at least 1,000 pairs of specific primers to generate a second set of amplification products. In various embodiments, at least 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different primer pairs are used.

In various embodiments of any aspect of the invention, the test primers are used to simultaneously amplify at least 1,000 different target loci in a sample containing DNA from the alleged father of the fetus, and also to simultaneously amplify target loci in samples containing maternal DNA from the pregnant mother of the fetus and fetal DNA to determine whether the alleged father is the biological father of the fetus.

In various embodiments of any aspect of the invention, the test primers are used to simultaneously amplify at least 1,000 different target loci in a cell or cells from an embryo to determine the presence or absence of chromosomal abnormalities. In various embodiments, cells from a set of two or more embryos are analyzed and one embryo is selected for in vitro fertilization.

In various embodiments of any aspect of the invention, the test primers are used to simultaneously amplify at least 1,000 different target loci in a forensic nucleic acid sample. In various embodiments, the length of the annealing step is greater than 3, 5, 8, 10, or 15 minutes.

In various embodiments of any aspect of the invention, the method includes simultaneously amplifying at least 1,000 different target loci in a control nucleic acid sample with test primers to generate a first set of target amplicons and simultaneously amplifying the target loci in a test nucleic acid sample to generate a second set of target amplicons, and comparing the first and second sets of target amplicons to determine whether a target locus is present in one sample but not in the other sample, or whether a target locus is present at different levels in the control sample and the test sample. In various embodiments, the test sample is from an individual suspected of having or at increased risk for a disease or phenotype of interest (e.g., cancer), and one or more target loci include a sequence (e.g., a polymorphism or other mutation) associated with an increased risk of or associated with the disease or phenotype of interest. In various embodiments, the method includes simultaneously amplifying 1,000 different target loci with test primers in a control sample comprising RNA to generate a first set of target amplicons and simultaneously amplifying the target loci in a test sample comprising RNA to generate a second set of target amplicons, and comparing the first and second sets of target amplicons to determine the presence or absence of differences in RNA expression levels between the control sample and the test sample. In various embodiments, the RNA is mRNA. In various embodiments, the test sample is from an individual suspected of having or at increased risk for a disease or phenotype of interest (e.g., cancer), and one or more target loci include a sequence (e.g., a polymorphism or other mutation) associated with an increased risk of or associated with the disease or phenotype of interest. In some embodiments, the test sample is from an individual diagnosed with a disease or phenotype of interest (e.g., cancer), where a difference in RNA expression levels between the control sample and the test sample indicates that the target locus contains a sequence (e.g., a polymorphism or other mutation) associated with an increased or decreased risk of the disease or phenotype of interest.

In some embodiments of any of the aspects of the invention, the test primers are selected from a library of candidate primers based on one or more parameters, such as the selection of primers using any of the methods of the invention. In some embodiments, the test primers are selected from the library of candidate primers based at least in part on the ability of the candidate primers to form primer dimers.

In one aspect, the invention features a method for selecting test primers from a library of candidate primers. In various embodiments, the selection includes (i) calculating an undesirability score for most or all possible combinations of two candidate primers from the library using a computer, where each undesirability score is based at least in part on the likelihood of dimer formation between the two candidate primers; (ii) removing the candidate primer from the library of candidate primers that has the highest undesirability score; (iii) if the candidate primer removed in step (ii) is a member of a primer pair, removing the other member of the primer pair from the library of candidate primers; and (iv) optionally repeating steps (ii) and (iii) to select a library of test primers. In some embodiments, the selection method is performed until all undesirability scores for the combinations of candidate primers remaining in the library are below a minimum threshold. In some embodiments, the selection method is performed until the number of candidate primers remaining in the library is reduced to a desired number. In various embodiments, the undesirability score is calculated for at least 80, 90, 95, 98, 99, or 99.5% of the possible combinations of candidate primers in the library. In various embodiments, the candidate primers remaining in the library are capable of simultaneously amplifying at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci. In various embodiments, the method also includes (v) contacting a nucleic acid sample comprising the target loci with the candidate primers remaining in the library to produce a reaction mixture; and (vi) subjecting the reaction mixture to primer extension reaction conditions to produce an amplification product comprising a target amplification product.

In one aspect, the invention features a method for selecting test primers from a library of candidate primers. In various embodiments, the method for selecting test primers from a library of candidate primers includes (i) calculating, using a computer, an undesirability score for most or all possible combinations of two candidate primers from the library, where each undesirability score is based at least in part on the likelihood of dimer formation between the two candidate primers; (ii) removing from the library of candidate primers a candidate primer that is part of the greatest number of combinations of two candidate primers that have an undesirability score that exceeds a first minimum threshold; (iii) if the candidate primer removed in step (ii) is a member of a primer pair, removing the remaining members of the primer pair from the library of candidate primers; and (iv) optionally repeating steps (ii) and (iii) to select a library of test primers. In some embodiments, the selection method is performed until the undesirability scores for the combinations of candidate primers remaining in the library are all equal to or less than the first minimum threshold. In some embodiments, the selection method is performed until the number of candidate primers remaining in the library is reduced to a desired number. In various embodiments, the undesirability score is calculated for at least 80, 90, 95, 98, 99, or 99.5% of the possible candidate primer combinations in the library. In various embodiments, the candidate primers remaining in the library are capable of simultaneously amplifying at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci. In various embodiments, the method also includes (v) contacting a nucleic acid sample comprising the target loci with the candidate primers remaining in the library to produce a reaction mixture; and (vi) subjecting the reaction mixture to primer extension reaction conditions to produce an amplification product comprising the target amplification product.

In various embodiments of any aspect of the invention, the selection method includes further reducing the number of candidate primers remaining in the library by lowering the first minimum threshold used in step (ii) to a lower second minimum threshold, and optionally repeating (ii) and (iii). In some embodiments, the selection method includes increasing the first minimum threshold used in step (ii) to a higher second minimum threshold, and optionally repeating (ii) and (iii). In some embodiments, the selection method is performed until all undesirability scores for the candidate primer combinations remaining in the library are equal to or less than the second minimum threshold, or until the number of candidate primers remaining in the library is reduced to a desired number.

In various embodiments of any aspect of the invention, the method includes, prior to step (i), identifying or selecting a primer that hybridizes to a target locus. In some embodiments, multiple primers (or primer pairs) hybridize to the same target locus, but one primer (or primer pair) is selected for the target locus using a selection method based on one or more parameters. In various embodiments, the method includes, prior to step (ii), removing from the library primer pairs that generate target amplicons that overlap with target amplicons generated by another primer pair. In various embodiments, candidate primers are selected from a group of two or more candidate primers with comparable undesirability scores based on one or more other parameters for removal from the candidate primer library. In some embodiments, the candidate primers remaining in the library are used as a test primer library in any method of the invention. In some embodiments, the resulting test primer library comprises any of the primer libraries of the invention.

In various embodiments of any aspect of the invention, the undesirability score is based at least in part on one or more parameters selected from the group consisting of the heterozygosity rate of the target locus, the prevalence associated with the sequence (e.g., polymorphism) of the target locus, the disease penetrance associated with the sequence (e.g., polymorphism) of the target locus, the specificity of the candidate primer for the target locus, the size of the candidate primer, the melting temperature of the target amplicon, the GC content of the target amplicon, the amplification efficiency of the target amplicon, and the size of the target amplicon.

In various embodiments of any of the aspects of the invention, the undesirability score is based at least in part on one or more parameters selected from the group consisting of the heterozygosity rate of the target locus, the specificity of the candidate primer for the target locus, the size of the candidate primer, the melting temperature of the target amplicon, the GC content of the target amplicon, the amplification efficiency of the target amplicon, and the size of the target amplicon, and the test primers are used to simultaneously amplify at least 1,000 different target loci in a sample that includes maternal DNA from the pregnant mother of the fetus and fetal DNA to determine the presence or absence of a fetal chromosomal abnormality. In various embodiments, the method includes ligating a universal primer binding site to DNA molecules in the sample, amplifying the ligated DNA molecules with at least 1,000 specific primers and a universal primer to generate a first set of amplification products, and amplifying the first set of amplification products with at least 1,000 pairs of specific primers to generate a second set of amplification products. In various embodiments, at least 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different primer pairs are used. In various embodiments, at least 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci are amplified.

In various embodiments of any aspect of the invention, the undesirability score is based at least in part on one or more parameters selected from the group consisting of the heterozygosity rate of the target locus, the specificity of the candidate primer for the target locus, the size of the candidate primer, the melting temperature of the target amplicon, the GC content of the target amplicon, the amplification efficiency of the target amplicon, and the size of the target amplicon, and the test primers are used to amplify at least 1,000 different target loci in a sample containing DNA from the alleged father of the fetus, and simultaneously amplify target loci in a sample containing maternal DNA from the pregnant mother of the fetus and fetal DNA to determine whether the alleged father is the biological father of the fetus. In various embodiments, at least 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci are amplified.

In various embodiments of any aspect of the invention, the undesirability score is based at least in part on one or more parameters selected from the group consisting of heterozygosity rate of the target locus, specificity of the candidate primer for the target locus, size of the candidate primer, melting temperature of the target amplicon, GC content of the target amplicon, amplification efficiency of the target amplicon, and size of the target amplicon, and the test primers are used to simultaneously amplify at least 1,000 different target loci in a cell or cells from an embryo to determine the presence or absence of a chromosomal abnormality. In various embodiments, cells from a set of two or more embryos are analyzed and one embryo is selected for in vitro fertilization. In various embodiments, at least 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci are amplified.

In various embodiments of any aspect of the invention, the undesirability score is based at least in part on one or more parameters selected from the group consisting of the heterozygosity rate of the target locus, the specificity of the candidate primer for the target locus, the size of the candidate primer, the melting temperature of the target amplicon, the GC content of the target amplicon, the amplification efficiency of the target amplicon, and the size of the target amplicon, and at least 1,000 different target loci in the forensic nucleic acid sample are simultaneously amplified using the test primers. In various embodiments, the length of the annealing step is greater than 3, 5, 8, 10, or 15 minutes. In various embodiments, at least 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci are amplified.

In various embodiments of any of the aspects of the invention, the undesirability score is based at least in part on one or more parameters selected from the group consisting of heterozygosity rate of the target locus, prevalence associated with the sequence (e.g., polymorphism) of the target locus, disease penetrance associated with the sequence (e.g., polymorphism) of the target locus, specificity of the candidate primer for the target locus, size of the candidate primer, melting temperature of the target amplicon, GC content of the target amplicon, amplification efficiency of the target amplicon, and size of the target amplicon, and the method includes simultaneously amplifying at least 1,000 different target loci in a control nucleic acid sample using the test primers to generate a first set of target amplicons and simultaneously amplifying the target loci in a test nucleic acid sample to generate a second set of target amplicons; and comparing the first and second sets of target amplicons to determine whether the target loci are present in one sample and absent in the other, or whether the target loci are present at different levels in the control sample and the test sample. In various embodiments, the test sample is from an individual suspected of having a disease or phenotype of interest, or an increased risk of a disease or phenotype of interest, where one or more target loci contain sequences (e.g., polymorphisms) associated with an increased risk of the disease or phenotype of interest at the target locus or associated with the disease or phenotype of interest. In various embodiments, at least 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci are amplified.

In various embodiments of any aspect of the invention, the undesirability score is based at least in part on one or more parameters selected from the group consisting of heterozygosity rate of the target locus, prevalence associated with the sequence (e.g., polymorphism) of the target locus, disease penetrance associated with the sequence (e.g., polymorphism) of the target locus, specificity of the candidate primer for the target locus, size of the candidate primer, melting temperature of the target amplicon, GC content of the target amplicon, amplification efficiency of the target amplicon, and size of the target amplicon, and the method includes simultaneously amplifying 1,000 different target loci in a control nucleic acid sample comprising RNA with the test primers to generate a first set of target amplicons, simultaneously amplifying the target loci in a test sample comprising RNA to generate a second set of target amplicons, and comparing the first and second sets of target amplicons to determine the presence or absence of differences in RNA expression levels between the control sample and the test sample. In various embodiments, the RNA is mRNA. In various embodiments, the test sample is from an individual suspected of having a disease or phenotype of interest (e.g., cancer) or an increased risk of a disease or phenotype of interest (e.g., cancer), where one or more target loci contain a sequence (e.g., a polymorphism or other mutation) associated with an increased risk of the disease or phenotype of interest or associated with the disease or phenotype of interest. In some embodiments, the test sample is from an individual diagnosed with a disease or phenotype of interest (e.g., cancer), where a difference in RNA expression levels between the control sample and the test sample indicates that the target locus contains a sequence (e.g., a polymorphism or other mutation) associated with an increase or decrease in the disease or phenotype of interest. In various embodiments, at least 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci are amplified.

In one aspect, the invention features a primer library. In some embodiments, primers are selected from a library of candidate primers using any of the methods of the invention. In some embodiments, the library includes primers that simultaneously hybridize to at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci. In some embodiments, the library includes primers that simultaneously amplify at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci. In some embodiments, the library contains primers that simultaneously amplify at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci, such that less than 60, 40, 30, 20, 10, 5, 4, 3, 2, 1, 0.5, 0.25, 0.1, or 0.05% of the amplification products are primer dimers. In some embodiments, the library comprises primers that simultaneously amplify 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci, such that at least 50, 60, 70, 80, 90, 95, 96, 97, 98, 99, or 99.5% of the amplified products are target amplified products. In some embodiments, the library comprises primers that simultaneously amplify target loci such that at least 50, 60, 70, 80, 90, 95, 96, 97, 98, 99, or 99.5% of the target loci from 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci are amplified. In some embodiments, the primer library includes at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 primer pairs, each primer pair including a forward test primer and a reverse test primer, each test primer pair hybridizing to a target locus. In some embodiments, the primer library includes at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 individual primers, each hybridizing to a different target locus, where the individual primers are not part of a primer pair.

In various embodiments of any aspect of the invention, the concentration of each primer is less than 100, 75, 50, 25, 10, 5, 2, or 1 nM. In various embodiments, the GC content of the primer is 30-80%, e.g., 40-70% or 50-60%. In some embodiments, the GC content range of the primer is less than 30, 20, 10, or 5%. In some embodiments, the melting temperature of the primer is 40-80° C., e.g., 50-70° C., 55-65° C., or 57-60.5° C. In some embodiments, the melting temperature range of the primer is less than 15, 10, 5, 3, or 1° C. In some embodiments, the length of the primer is 15-100 nucleotides, e.g., 15-75 nucleotides, 15-40 nucleotides, 17-35 nucleotides, 18-30 nucleotides, or 20-65 nucleotides in length. In some embodiments, the primer comprises a tag that is not target specific, e.g., a tag that forms an internal loop structure. In some embodiments, the tag is between two DNA binding regions. In various embodiments, the primer comprises a 5' region that is specific for the target locus, an internal region that is not specific for the target locus and forms a loop structure, and a 3' region that is specific for the target locus. In various embodiments, the length of the 3' region is at least 7 nucleotides. In some embodiments, the length of the 3' region is 7-20 nucleotides, e.g., 7-15 nucleotides, or 7-10 nucleotides. In various embodiments, the primer comprises a 5' region that is not specific for the target locus (e.g., another tag or a universal primer binding site), followed by a region that is specific for the target locus, an internal region that is not specific for the target locus and forms a loop structure, and a 3' region that is specific for the target locus. In some embodiments, the length of the primer ranges from less than 50, 40, 30, 20, 10, or 5 nucleotides. In some embodiments, the length of the target amplicon is 50-100 nucleotides, e.g., 60-80 nucleotides, or 60-75 nucleotides. In some embodiments, the length range of the target amplicon is less than 50, 25, 15, 10, or 5 nucleotides.

In one aspect, the invention provides a kit comprising any of the primer libraries of the invention for amplifying a target locus in a nucleic acid sample. In some embodiments, the kit includes instructions for using the library to amplify the target locus.

In one aspect, the invention features a method for determining the ploidy state of chromosomes of a gestating fetus. In some embodiments, the method includes contacting a nucleic acid sample with a primer library that simultaneously hybridizes to at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 distinct polymorphic loci to generate a reaction mixture, wherein the nucleic acid sample includes maternal DNA from the mother of the fetus and fetal DNA from the fetus. In some embodiments, the reaction mixture is subjected to primer extension reaction conditions to generate amplification products, sequencing data is generated from the amplification products using a high-throughput sequencer, the number of alleles at the polymorphic locus is calculated by a computer based on the sequencing data, a plurality of ploidy hypotheses, each associated with a different possible ploidy state of the chromosome, a joint distribution model is constructed by a computer for the number of alleles predicted at the polymorphic locus of the chromosome for each ploidy hypothesis, the relative probability of each ploidy hypothesis is calculated by a computer using the joint distribution model and the number of alleles, and the ploidy state of the fetus is called by selecting the ploidy state corresponding to the hypothesis with the greatest probability.

In one aspect, the invention features a method for determining the ploidy state of a chromosome in a gestating fetus. In one embodiment, a method for determining the ploidy state of a chromosome in a gestating fetus includes obtaining a first sample of DNA including maternal DNA from the mother of the fetus and fetal DNA from the fetus, preparing the first sample by isolating the DNA to obtain a prepared sample, measuring the DNA in the prepared sample at a plurality of polymorphic loci on the chromosome, calculating by a computer the number of alleles at the plurality of polymorphic loci from the DNA measurements made on the prepared sample, generating by a computer a plurality of ploidy hypotheses, each of which relates to a different possible ploidy state of the chromosome, constructing by a computer a joint distribution model for the expected number of alleles at the plurality of polymorphic loci on the chromosome for each ploidy hypothesis, determining by a computer the relative probability of each of the ploidy hypotheses using the joint distribution model and the number of alleles measured in the prepared sample, and calling the ploidy state of the fetus by selecting the ploidy state corresponding to the hypothesis with the greatest probability.

In one aspect, the invention features a method for testing for abnormal chromosomal distribution in a sample that includes a mixture of maternal and fetal DNA. In some embodiments, the method includes the steps of (i) contacting the sample with a primer library that simultaneously hybridizes to at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci to generate a reaction mixture, the target loci being from a plurality of different chromosomes, the plurality of different chromosomes including at least one first chromosome suspected of having an abnormal distribution in the sample and at least one second chromosome presumed to be normally distributed in the sample; and (ii) contacting the sample with a primer library that simultaneously hybridizes to at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci to generate a reaction mixture, the target loci being from a plurality of different chromosomes, the plurality of different chromosomes including at least one first chromosome suspected of having an abnormal distribution in the sample and at least one second chromosome presumed to be normally distributed in the sample. (iii) subjecting the mixture to primer extension reaction conditions to produce an amplification product; (iii) sequencing the amplification product to obtain a plurality of sequence tags aligned to the target loci, the sequence tags being of sufficient length to be assigned to specific target loci; (iv) assigning, by a computer, the plurality of sequence tags to their corresponding target loci; (v) determining, by a computer, a number of sequence tags to be assigned to the target loci of the first chromosome and a number of sequence tags to be assigned to the target loci of the second chromosome; and (vi) comparing, by a computer, the numbers from step (v) to determine the presence or absence of an abnormal distribution of the first chromosome.

In one aspect, the present invention provides a method for detecting the presence or absence of fetal aneuploidy. In some embodiments, the method includes: (i) contacting a sample comprising a mixture of maternal and fetal DNA with a library of primers that simultaneously hybridize to at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different non-polymorphic target loci from a plurality of different chromosomes to generate a reaction mixture; (ii) subjecting the reaction mixture to primer extension reaction conditions to generate an amplification product comprising a target amplicon; (iii) quantifying on a computer the relative frequency of the target amplicons from the first and second chromosomes of interest; (iv) comparing on a computer the relative frequency of the target amplicons from the first and second chromosomes of interest; and (v) identifying the presence or absence of an aneuploidy based on the compared relative frequency of the first and second chromosomes of interest. In some embodiments, the first chromosome is a chromosome suspected to be euploid. In some embodiments, the second chromosome is a chromosome suspected to be aneuploid.

In one aspect, a method for determining the presence or absence of fetal aneuploidy in a maternal tissue sample comprising fetal genomic DNA and maternal genomic DNA, comprising the steps of: (a) obtaining a mixture of fetal genomic DNA and maternal genomic DNA from the maternal tissue sample; (b) performing massively parallel DNA sequencing of DNA fragments randomly selected from the mixture of fetal genomic DNA and maternal genomic DNA of step (a) to determine sequences of the DNA fragments; (c) identifying the chromosome to which the sequence obtained in step (b) belongs; (d) using the data of step (c) to determine the amount of at least one first chromosome in the mixture of maternal genomic DNA and fetal genomic DNA, the at least one first chromosome being presumed to be euploid in the fetus; and (e) using the data of step (c) to determine the amount of at least one first chromosome in the mixture of maternal genomic DNA and fetal genomic DNA. The method includes the steps of: (f) determining the amount of a second chromosome, the second chromosome being suspected to be aneuploid in the fetus; (g) calculating the percentage of fetal DNA in the mixture of fetal DNA and maternal DNA; (h) calculating the predicted distribution of the amount of the second target chromosome using the number of step (d) if the second target chromosome is euploid; and (i) determining whether the amount of the second chromosome determined in step (e) is more likely to be part of the distribution calculated in step (g) or the distribution calculated in step (h), using a maximum likelihood or maximum a posteriori method, thereby indicating the presence or absence of fetal aneuploidy.

In various embodiments of any of the aspects of the invention, the method also includes obtaining genotype data from one or both parents of the fetus. In some embodiments, obtaining genotype data from one or both parents of the fetus includes preparing DNA from the parents, the DNA preferentially enriching for DNA at a plurality of polymorphic loci to obtain prepared parental DNA, optionally amplifying the prepared parental DNA, and measuring the parental DNA in the prepared sample at the plurality of polymorphic loci.

In various embodiments of any of the aspects of the invention, constructing a joint distribution model for the expected allele count probabilities at multiple polymorphic loci on a chromosome is performed using genetic data obtained from one or both parents. In some embodiments, a sample (e.g., a first sample) is isolated from maternal plasma, and obtaining genotype data from the mother is performed by estimating the maternal genotype data from DNA measurements made on the prepared sample.

In one aspect, a diagnostic box is disclosed that is useful for determining the ploidy status of chromosomes in a gestational fetus, the diagnostic box being capable of carrying out any of the preparation and measurement steps of the methods of the present invention.

In various embodiments of any of the aspects of the invention, the allele counts are probabilistic rather than binary. In some embodiments, measurements of DNA in the prepared sample at multiple polymorphic loci are also used to determine whether the fetus has one or more disease-linked haplotypes.

In various embodiments of any of the aspects of the invention, the step of constructing a joint distribution model for allele count probabilities is performed by modeling the dependency between polymorphic alleles on a chromosome using data on the probability of chromosomal crossover at different locations within the chromosome. In some embodiments, the steps of constructing a joint distribution model for allele counts and determining the relative probability of each hypothesis are performed using a method that does not require the use of a reference chromosome.

In various embodiments of any aspect of the invention, the step of determining the relative probability of each hypothesis uses an estimated fraction of fetal DNA in the prepared sample. In some embodiments, the DNA measurements from the prepared sample used in the steps of calculating the allele count probabilities and determining the relative probability of each hypothesis comprise primary genetic data. In some embodiments, the step of selecting the ploidy state corresponding to the hypothesis with the greatest probability is performed using maximum likelihood estimation or maximum a posteriori estimation.

In various embodiments of any of the aspects of the invention, calling the ploidy state of the fetus also includes combining the relative probabilities of each of the ploidy hypotheses determined using a joint distribution model and allele count probabilities with the relative probabilities of each of the ploidy hypotheses calculated using a statistical technique selected from the group consisting of read count analysis, comparison of heterozygosity rates, statistics available only using parental genetic information, genotype signal probabilities normalized to a particular parental status, statistics calculated using an estimated fetal fraction of a sample (e.g., a first sample) or prepared sample, and combinations thereof.

In various embodiments of any of the aspects of the invention, a confidence estimate is calculated for the called ploidy state. In some embodiments, the method also includes taking a clinical action selected from one of terminating the pregnancy or maintaining the pregnancy based on the called fetal ploidy state.

In various embodiments of any aspect of the invention, the method may be performed on a fetus between 4 and 5 weeks of gestation; between 5 and 6 weeks of gestation; between 6 and 7 weeks of gestation; between 7 and 8 weeks of gestation; between 8 and 9 weeks of gestation; between 9 and 10 weeks of gestation; between 10 and 12 weeks of gestation; between 12 and 14 weeks of gestation; between 14 and 20 weeks of gestation; between 20 and 40 weeks of gestation; early pregnancy; mid pregnancy; late pregnancy; or combinations thereof.

In various embodiments of any of the aspects of the invention, the method is used to generate a report showing the ploidy state of the determined chromosome in a gestating fetus. In some embodiments, a kit for determining the ploidy state of a target chromosome in a gestating fetus designed for use with any of the methods of the invention is disclosed, the kit including a plurality of inner forward primers and, optionally, a plurality of inner reverse primers, each of which is designed to hybridize to a region of DNA immediately upstream and/or downstream of one of the polymorphic sites on the target chromosome, and, optionally, an additional chromosome, the hybridizing region of which is separated from the polymorphic site by a small number of bases, the small number being selected from the group consisting of 1, 2, 3, 4, 5, 6-10, 11-15, 16-20, 21-25, 26-30, 31-60, and combinations thereof.

In one aspect, the invention features a method for determining whether an alleged father is the biological father of a fetus carried by a pregnant mother. In some embodiments, the method includes the steps of: (i) simultaneously amplifying a plurality of polymorphic loci, including polymorphic loci on at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different genetic material from the alleged father to generate a first set of amplified products; (ii) simultaneously amplifying a corresponding plurality of polymorphic loci in a mixed sample of DNA, including fetal DNA and maternal DNA, from a blood sample of the pregnant mother to generate a second set of amplified products; (iii) determining, on a computer, a probability that the alleged father is the biological father of the fetus using genotypic measurements based on the first and second sets of amplified products; and (iv) establishing whether the alleged father is the biological father of the fetus using the determined probability that the alleged father is the biological father of the fetus. In various embodiments, the method further includes simultaneously amplifying polymorphic loci on corresponding genetic material from the mother to generate a third set of amplification products, where the probability that the alleged father is the biological father of the fetus is determined using genotype measurements based on the first, second, and third sets of amplification products.

In one aspect, the invention provides a method for estimating the relative likelihood of each embryo from a set of embryos to develop as desired. In some embodiments, the method includes contacting a sample from each embryo with a library of primers that simultaneously hybridize to at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci to generate a reaction mixture for each embryo, the sample being each obtained from a cell from one or more embryos. In some embodiments, each reaction mixture is subjected to primer extension reaction conditions to generate an amplification product. In some embodiments, the method includes determining, in a computer, one or more characteristics of at least one cell from each embryo based on the amplification product, and estimating, in a computer, the relative likelihood of each embryo to develop as desired based on the one or more characteristics of the at least one cell of each embryo.

In one aspect, the invention features a method for measuring the amount of two or more target loci in a nucleic acid sample. In some embodiments, the method includes (i) using PCR to amplify a nucleic acid sample that includes a first reference locus, a second reference locus, a first target locus, and a second target locus to form an amplification product, where the first reference locus and the first target locus have the same number of nucleotides but have a different sequence at one or more nucleotide positions, and the second reference locus and the second target locus have the same number of nucleotides but have a different sequence at one or more nucleotide positions; and (ii) sequencing the amplification product to determine the amount of the amplified product. (i) determining a reference ratio comparing the relative amount of the amplified first target locus compared to the amplified second target locus, where the reference ratio indicates a difference in PCR efficiency for amplification of the first reference locus and the second reference locus, (iii) determining a target ratio comparing the relative amount of the amplified first target locus compared to the amplified second target locus, and (iv) adjusting the target ratio from step (iii) based on the reference ratio from step (ii) to determine the relative amounts of the first target locus and the second target locus in the sample. In various embodiments, the method includes determining the absolute amounts of the first target locus and the second target locus in the sample. In various embodiments, the method further comprises determining the presence or absence of target loci (e.g., at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci) in the sample. In various embodiments, the method comprises using any of the primer libraries of the invention. In various embodiments, the method comprises simultaneously amplifying 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci.

In one aspect, the invention features a method for quantitatively measuring multiple genetic targets in an analytical sample. In some embodiments, the method includes (i) mixing genetic material from the analytical sample with multiple target-specific amplification reagents and multiple reference sequences corresponding to the targets of the target-specific amplification reagents, (ii) amplifying target regions of the genetic material and the reference sequences to generate target amplicons and reference sequence amplicons, and (iii) measuring the amount of the generated target amplicons and reference sequence amplicons. In some embodiments, the genetic material is present in a genetic library. In some embodiments, the genetic targets are polymorphic loci (such as SNPs). In some embodiments, measuring the amount is achieved by counting sequences. In some embodiments, the method further includes measuring an estimated copy number of at least one chromosome in the sample from which the genetic library is derived, the measuring comprising comparing the number of sequence reads of the target amplicons to the number of sequence reads of the reference amplicons. In some embodiments, the reference sequences and the genetic library include universal priming sites that can be primed by the same primer. In some embodiments, the mixing step comprises at least 10; 100; 500; 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target-specific amplification reagents and at least 10; 100; 500; 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 standard sequences. In various embodiments, the method comprises using any of the primer libraries of the invention. In various embodiments, the method includes simultaneously amplifying 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target regions. In some embodiments, the relative amount of each of the standard sequences is known. In some embodiments, the relative amount of each of the sequences is calibrated to a reference genome. In some embodiments, the analytical sample includes a mixture of fetal and maternal genomes. In some embodiments, the analytical sample is derived from the blood or plasma of a pregnant woman. In some embodiments, the reference genome has at least one aneuploidy, such as an aneuploidy at chromosome 13, 18, 21, X, or Y. In some embodiments, the reference genome is diploid.

In one aspect, the invention features a mixture comprising a plurality of gene standard sequences, the relative amount of each gene standard sequence in the mixture being determined by calibration to a reference genome. In various embodiments, the mixture comprises at least 10; 100; 500; 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 gene standard sequences. In various embodiments, the gene standard sequence comprises a first universal priming site, a second universal priming site, a first target specific priming site, a second target specific priming site, and a marker sequence located between the first and second target specific priming sites, the first target specific site and the second target specific priming site being located between the first and second universal priming sites. In various embodiments, the calibration comprises using any of the primer libraries of the present invention. In various embodiments, the calibration comprises simultaneously amplifying 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target regions. In some embodiments, the reference genome has at least one aneuploidy, such as an aneuploidy at chromosome 13, 18, 21, X, or Y. In some embodiments, the reference genome is diploid.

In one aspect, the invention features a method for generating a set of calibrated genetic standard sequences. In some embodiments, the method includes (i) forming an amplification reaction mixture including a genetic library prepared from a reference genome, a set of target-specific amplification primer reagents, and a set of genetic standard sequences corresponding to the set of target-specific amplification reagents; (ii) amplifying the genetic library and the genetic standard sequences to generate target sequence-derived amplification products and genetic standard sequence-derived amplification products; (iii) measuring the amount of the target sequence-derived amplification products and the genetic standard sequence-derived amplification products; and (iv) determining the relative amounts of each of the genetic standard sequences to each other, thereby calibrating the multiple genetic standard sequences. In various embodiments, at least 10; 100; 500; 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 genetic reference sequences are used. In various embodiments, the method includes using any of the primer libraries of the invention. In various embodiments, the invention includes simultaneously amplifying 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different sequences. In some embodiments, the reference genome has at least one aneuploidy, such as an aneuploidy at chromosome 13, 18, 21, X, or Y. In some embodiments, the reference genome is diploid.

In one aspect, the invention provides a set of genetic reference sequences that have been calibrated by any of the methods of the invention. In one aspect, the invention provides a set of genetic reference sequences that can be calibrated before, during, or after the method is performed.

In one aspect, the invention features a method for determining the copy number of a gene of interest that includes at least one allele having a deletion. In some embodiments, the method includes (i) mixing genetic material from a sample for analysis with an amplification reagent specific for the gene of interest and unable to significantly amplify the deletion including the allele of the gene of interest, a standard sequence corresponding to the gene of interest, an amplification reagent specific for the reference sequence, and a standard sequence corresponding to the reference sequence; (ii) amplifying the gene sequence of interest, the standard sequence corresponding to the gene of interest, the reference sequence, and the standard sequence corresponding to the reference sequence to generate a gene of interest amplicon, a reference sequence amplicon, and a standard sequence amplicon; and (iii) measuring the amount of the generated target amplicon and standard sequence amplicon. In some embodiments, the measuring of the amount is achieved by counting sequence reads. In some embodiments, the method further includes calculating an estimated copy number of a chromosome in a sample from which at least one genetic library is derived, the calculating step including a step of comparing the number of sequences of the target amplicon to the number of sequences of the standard amplicon. In some embodiments, the standard sequence and the genetic library include universal priming sites that can be primed by the same primer. In some embodiments, the relative amount of each sequence is calibrated against a reference genome. In various embodiments, at least 10; 100; 500; 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 gene reference sequences are used. In various embodiments, the method includes using any of the primer libraries of the present invention. In various embodiments, the method includes simultaneously amplifying 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target regions. In some embodiments, the reference genome is diploid. In some embodiments, the analytical sample is derived from blood.

In various embodiments of any of the aspects of the invention, the step of preferentially enriching DNA in a sample (e.g., a first sample) at target loci (e.g., a plurality of polymorphic loci) includes using a plurality of pre-circularization probes, each of which targets one of the loci (e.g., polymorphic loci), and preferably designed such that the 3' and 5' ends of the probe hybridize to a region of DNA separated from the polymorphic site of the locus by a small number of bases, the small number being 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 9 The method includes the steps of obtaining a probe having a length of 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21-25, 26-30, 31-60, or a combination thereof, hybridizing the pre-circularized probe with DNA from a sample (e.g., a first sample), filling the gap between the hybridized probe ends with a DNA polymerase, circularizing the pre-circularized probe, and amplifying the circularized probe.

In various embodiments of any of the aspects of the invention, the step of preferentially enriching DNA at a target locus (e.g., a plurality of polymorphic loci) includes using a plurality of ligation-mediated PCR probes, each PCR probe targeting one of the target loci (e.g., polymorphic loci), with upstream and downstream PCR probes designed to hybridize to a region of DNA on one strand of the DNA that is separated from the polymorphic site of the locus by a small number of bases, preferably, the small number being 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, The method includes obtaining a PCR probe that is 13, 14, 15, 16, 17, 18, 19, 20, 21-25, 26-30, 31-60, or a combination thereof, hybridizing the ligation-mediated PCR probe with DNA from a sample (e.g., a first sample), filling the gap between the ends of the ligation-mediated PCR probe with a DNA polymerase, ligating the ligation-mediated PCR probe, and amplifying the ligated ligation-mediated PCR probe.

In some embodiments of various aspects of the invention, preferentially enriching DNA at target loci (e.g., multiple polymorphic loci) includes obtaining multiple hybrid capture probes that target the loci (e.g., multiple polymorphic loci), hybridizing the hybrid capture probes to DNA in a sample (e.g., a first sample), and physically removing some or all of the unhybridized DNA from the DNA-related sample (e.g., the first sample).

In some embodiments of any aspect of the invention, the hybrid capture probe is designed to hybridize to a region adjacent to but not overlapping with the polymorphic site. In some embodiments, the hybrid capture probe is designed to hybridize to a region adjacent to but not overlapping with the polymorphic site, and the length of the adjacent capture probe can be selected from the group consisting of less than about 120 bases, less than about 110 bases, less than about 100 bases, less than about 90 bases, less than about 80 bases, less than about 70 bases, less than about 60 bases, less than about 50 bases, less than about 40 bases, less than about 30 bases, and less than about 25 bases. In some embodiments, the hybrid capture probe is designed to hybridize to a region overlapping with the polymorphic site, and the plurality of hybrid capture probes includes at least two hybrid capture probes for each polymorphic locus, each hybrid capture probe designed to be complementary to another allele at one polymorphic locus.

In some embodiments of any of the aspects of the invention, the step of preferentially enriching DNA at a plurality of polymorphic loci includes the steps of obtaining a plurality of inner forward primers, each of which targets one of the polymorphic loci and the 3' end of the inner forward primer is designed to hybridize to a region of DNA upstream of the polymorphic site and separated from the polymorphic site by a small number of bases, the small number being selected from the group consisting of 1 base pair, 2 base pairs, 3 base pairs, 4 base pairs, 5 base pairs, 6-10 base pairs, 11-15 base pairs, 16-20 base pairs, 21-25 base pairs, 26-30 base pairs, or 31-60 base pairs; and, optionally, obtaining a plurality of inner forward primers, each of which targets one of the polymorphic loci and the 3' end of the inner forward primer is designed to hybridize to a region of DNA upstream of the polymorphic site and separated from the polymorphic site by a small number of bases, the small number being selected from the group consisting of 1 base pair, 2 base pairs, 3 base pairs, 4 base pairs, 5 base pairs, 6-10 base pairs, 11-15 base pairs, 16-20 base pairs, 21-25 base pairs, 26-30 base pairs, or 31-60 base pairs. The method includes obtaining a primer having a 3' end and a reverse primer on each side, each of which targets one of the polymorphic loci, the 3' end of the inner reverse primer being designed to hybridize to a region of DNA that is upstream of the polymorphic site and separated from the polymorphic site by a small number of bases, the small number being selected from the group consisting of 1 base pair, 2 base pairs, 3 base pairs, 4 base pairs, 5 base pairs, 6-10 base pairs, 11-15 base pairs, 16-20 base pairs, 21-25 base pairs, 26-30 base pairs, or 31-60 base pairs, hybridizing the inner primer to the DNA, and amplifying the DNA using the polymerase chain reaction to form an amplification product.

In some embodiments of any of the aspects of the invention, the method further includes obtaining a plurality of outer forward primers, each of which targets one of the target loci (e.g., polymorphic loci) and is designed to hybridize with a region of DNA upstream of the inner forward primer, and optionally obtaining a plurality of outer reverse primers, each of which targets one of the target loci (e.g., polymorphic loci) and is designed to hybridize with a region of DNA immediately downstream of the inner reverse primer, hybridizing the first primer to the DNA, and amplifying the DNA using polymerase chain reaction.

In some embodiments of any of the aspects of the invention, the method further includes obtaining a plurality of outer reverse primers, each of which targets one of the target loci (e.g., polymorphic loci) and is designed to hybridize with a region of DNA immediately downstream of the inner reverse primer, and optionally obtaining a plurality of outer forward primers, each of which targets one of the polymorphic loci and is designed to hybridize with a region of DNA upstream of the inner forward primer, hybridizing the first primer to the DNA, and amplifying the DNA using polymerase chain reaction.

In some embodiments of any aspect of the invention, the sample (e.g., the first sample) preparation step further includes adding universal adaptors to DNA in the sample (e.g., the first sample) and amplifying the DNA in the sample (e.g., the first sample) using polymerase chain reaction. In some embodiments, at least a portion of the amplified amplicons is less than 100 bp, less than 90 bp, less than 80 bp, less than 70 bp, less than 65 bp, less than 60 bp, less than 55 bp, less than 50 bp, or less than 45 bp, where the portion is 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or 99%.

In some embodiments of any aspect of the invention, the step of amplifying the DNA is performed in one or more individual reaction volumes, each of which contains more than 100 different forward and reverse primer pairs, more than 200 different forward and reverse primer pairs, more than 500 different forward and reverse primer pairs, more than 1,000 different forward and reverse primer pairs, more than 2,000 different forward and reverse primer pairs, more than 5,000 different forward and reverse primer pairs, more than 10,000 different forward and reverse primer pairs, more than 20,000 different forward and reverse primer pairs, more than 50,000 different forward and reverse primer pairs, or more than 100,000 different forward and reverse primer pairs.

In some embodiments of any aspect of the invention, preparing the sample (e.g., the first sample) further comprises dividing the sample (e.g., the first sample) into a plurality of portions, with DNA in each portion preferentially enriched at a subset of the target loci (e.g., a plurality of polymorphic loci). In some embodiments, the inner primers are selected by identifying primer pairs that may form undesired primer duplexes and removing at least one of the pairs of primers identified as likely to form undesired primer duplexes from the plurality of primers. In some embodiments, the inner primers contain a region designed to hybridize either upstream or downstream of the target locus (e.g., the polymorphic locus) and, optionally, a universal priming sequence designed to enable PCR amplification. In some embodiments, at least some of the primers further contain a random region that is different for each individual primer molecule. In some embodiments, at least some of the primers further contain a molecular barcode.

In some embodiments of any aspect of the invention, the preferential enrichment results in an average allelic bias between the prepared sample and the sample (e.g., the first sample) of a magnitude selected from the group consisting of 2-fold or less, 1.5-fold or less, 1.2-fold or less, 1.1-fold or less, 1.05-fold or less, 1.02-fold or less, 1.01-fold or less, 1.005-fold or less, 1.002-fold or less, 1.001-fold or less, and 1.0001-fold or less. In some embodiments, the plurality of polymorphic loci are SNPs. In some embodiments, the step of measuring DNA in the prepared sample is performed by sequencing.

In some embodiments of any of the aspects of the invention, the target loci are present on the same nucleic acid (e.g., the same chromosome or the same region of a chromosome) of the control. In some embodiments, at least some of the target loci are present on different nucleic acids (e.g., different chromosomes) of the subject. In some embodiments, the nucleic acid sample comprises fragmented or digested nucleic acid. In some embodiments, the nucleic acid sample comprises genomic DNA, cDNA, or mRNA. In some embodiments, the nucleic acid sample comprises DNA from a single cell. In some embodiments, the nucleic acid sample is a substantially cell-free blood or plasma sample. In some embodiments, the nucleic acid sample comprises or is derived from blood, plasma, saliva, semen, sperm, cell culture supernatant, mucus secretion, dental plaque, gastrointestinal tissue, stool, urine, hair, bone, bodily fluids, tears, tissue, skin, nail, blastomere, embryo, amniotic fluid, chorionic villus sample, bile, lymph, cervical mucus, or a forensic sample. In some embodiments, the target loci are segments of human nucleic acid. In some embodiments, the target locus comprises or consists of a single nucleotide polymorphism (SNP). In some embodiments, the primer is a DNA molecule.

In some embodiments of any of the aspects of the invention, the DNA in the sample (e.g., the first sample) is derived from maternal plasma. In some embodiments, preparing the sample (e.g., the first sample) further comprises amplifying the DNA. In some embodiments, preparing the sample (e.g., the first sample) further comprises preferentially enriching the DNA in the sample (e.g., the first sample) at target loci (e.g., multiple polymorphic loci).

In various embodiments, the primer extension reaction or polymerase chain reaction includes the addition of one or more nucleotides by a polymerase. In various embodiments, the primer extension reaction or polymerase chain reaction does not include ligation-mediated PCR. In various embodiments, the primer extension reaction or polymerase chain reaction does not include ligation of two primers by a ligase. In various embodiments, the primer does not include a ligated inverse probe (LIP), which is also referred to as a pre-circularized probe, pre-circularizing probe or circularizing probe, circularizing probe, Padlock probe, or molecular inversion probe (MIP).

All aspects and embodiments of the invention described herein will be understood to include "comprising," "consisting," and "consisting essentially of" aspects and embodiments.

Definition Single nucleotide polymorphism (SNP) refers to a single base that may differ between the genomes of two members of the same species. The use of this term should not imply any limitation on the frequency with which each variant occurs.

Sequence refers to a DNA sequence or a gene sequence. Sequence may refer to the primary physical structure of an individual's DNA molecule or strand. Sequence may refer to the sequence of nucleotides found in a DNA molecule or a complementary strand of a DNA molecule. Sequence may refer to the information contained in a DNA molecule, represented in silico.

A locus refers to a specific region of interest on an individual's DNA and may refer to a SNP, which is a site of possible insertion or deletion or some other relevant genetic variation. A disease-linked SNP may also refer to a disease-linked locus.

Polymorphic allele, also "polymorphic locus", refers to an allele or locus whose genotype varies among individuals within a given species. Some examples of polymorphic alleles include single nucleotide polymorphisms, short tandem repeats, deletions, duplications, and inversions.

A polymorphic site refers to a specific nucleotide found in a polymorphic region that varies between individuals.

An allele refers to a gene that occupies a particular locus.

Genetic data, also "genotypic data", refers to data describing aspects of the genome of one or more individuals. It may refer to a locus or set of loci, a partial or full sequence, a portion or entire chromosome, or the entire genome. It may refer to the identity of one or more nucleotides, it may refer to a set of sequential nucleotides or nucleotides from different locations in the genome, or a combination thereof. Genotypic data is generally in silico, but it is also possible to think of the physical nucleotides in a sequence as chemically encoded genetic data. Genotypic data can be said to be "on", "of", "at", "from" or "on" an individual(s). Genotypic data may refer to output measurements from a genotyping platform when these measurements are made on genetic material.

Genetic material, also "genetic sample", refers to physical material, such as tissue or blood, from one or more individuals that contains DNA or RNA.

Noisy genetic data refers to genetic data that has any of the following: allele dropouts, uncertain base pair measurements, inaccurate base pair measurements, missing base pair measurements, uncertain measurements of insertions or deletions, uncertain measurements of chromosomal segment copy numbers, false signals, missing measurements, other errors, or combinations thereof.

Confidence refers to the statistical likelihood that a called SNP, allele, set of alleles, ploidy call, or determined number of chromosome segment copies accurately represents an individual's true genetic status.

Ploidy calling, also "chromosome copy number calling" or "copy number calling" (CNC), can refer to the act of determining the amount and/or identity of one or more chromosomes present in a cell.

Aneuploidy refers to the condition in which there is an incorrect number of chromosomes in a cell (e.g., the presence of an incorrect number of complete chromosomes or an incorrect number of chromosome segments, e.g., a deletion or duplication of a chromosome segment). In the case of human somatic cells, aneuploidy can refer to when a cell does not contain the 22 pairs of autosomes and one pair of sex chromosomes. In the case of human gametes, aneuploidy can refer to when a cell does not contain one of each of the 23 chromosomes. In the case of monochromosomal types, aneuploidy can refer to when there are roughly two homologous but not identical copies of a chromosome, or when there are two copies of a chromosome that originate from the same parent. In some embodiments, the deletion of a chromosome segment is a microdeletion.

Ploidy state refers to the amount and/or chromosomal identity of one or more chromosome types in a cell.

Chromosome may refer to a single chromosome copy, meaning a single molecule of DNA present in normal somatic cells in 46 copies, an example of which is "maternally derived chromosome 18". Chromosome may also refer to a chromosome type present in normal human somatic cells in 23 copies, an example of which is "chromosome 18".

Chromosomal identity can refer to the number of chromosomes referred to, i.e., chromosome type. A normal human has 22 numbered autosomal types and two sex chromosomes. Chromosomal identity can also refer to the parental origin of the chromosome. Chromosomal identity can also refer to the specific chromosome inherited from a parent. Chromosomal identity can also refer to other identifying features of a chromosome.

The state of genetic material or simply "genetic state" can refer to the identity of a set of SNPs on DNA, the haplotype of phased genetic material, and the sequence of DNA, including insertions, deletions, repeats, and mutations. It can also refer to the ploidy state of one or more chromosomes, a chromosome segment, or a set of chromosome segments.

Allele data refers to a collection of genotype data for a set of one or more alleles. Allele data may refer to phased haplotype data. Allele data may refer to SNP identities, and allele data may refer to DNA sequence data including insertions, deletions, repeats, and mutations. Allele data may encompass each allele of parental origin.

An allele state refers to the actual state of a gene within a set of one or more alleles. An allele state may refer to the actual state of a gene as described in the allele data.

Allelic ratio or allele ratio refers to the ratio between the amount of each allele at a locus present in a sample or individual. If the sample is measured by sequencing, the allelic ratio may refer to the ratio of sequence reads that map to each allele at a locus. If the sample is measured by an intensity-based measurement method, the allelic ratio may refer to the ratio of the amount of each allele present at the locus as estimated by the measurement method.

Allele count refers to the number of sequences that map to a particular locus; if the locus is polymorphic, the allele count refers to the number of sequences that map to each of the alleles. If each allele is counted in a binary fashion, the allele count will be an integer. If alleles are counted probabilistically, the allele count can be a fraction.

Allele count probability refers to the number of sequences that may map to a set of alleles at a particular locus or polymorphic locus combined with the probability of mapping. Note that the allele count, where the probability of mapping for each of the counted sequences is binary (0 or 1), is equal to the allele count probability. In some embodiments, the allele count probability may be binary. In some embodiments, the allele count probability may be set equal to the DNA measurement.

Allele distribution or "allele count distribution" refers to the relative amount of each allele present at each locus in a set of loci. Allele distribution may refer to an individual, a sample, or a set of measurements taken on a sample. In the context of sequencing, allele distribution refers to the number of reads or likelihoods that map to a particular allele for each allele in a set of polymorphic loci. Allele measurements can be treated probabilistically, i.e., the likelihood that a given allele will represent a given sequence read is a fraction between 0 and 1, or allele measurements can be treated in a binary manner, i.e., any given read is considered to be exactly 0 or 1 copies of a particular allele.

An allele distribution pattern refers to a set of different allele distributions for different parental situations. A particular allele distribution pattern may indicate a particular ploidy state.

Allelic bias refers to the degree to which the measured ratio of alleles at a heterozygous locus differs from the ratio present in the original DNA sample. The degree of allelic bias at a particular locus is equal to the measured allele ratio at that locus divided by the ratio of alleles in the original DNA sample at that locus. Allelic bias can be defined as greater than 1, so if a calculation of the degree of allelic bias results in a value x less than 1, the degree of allelic bias can be rephrased as 1/x. Allelic bias can be due to amplification bias, purification bias, or some other phenomenon that affects different alleles differently.

A primer, also a "PCR probe," refers to a single DNA molecule (DNA oligomer) or a population of DNA molecules (DNA oligomers), where the DNA molecules are identical or nearly identical, and the primer contains a region designed to hybridize with a target locus (e.g., a target polymorphic or non-polymorphic locus) and may contain a priming sequence designed to enable PCR amplification. The primer may also contain a molecular barcode. The primer may contain a random region that is different for each individual molecule. The terms "test primer" and "candidate primer" are not meant to be limiting and may refer to any of the primers disclosed herein.

A primer library refers to a population of two or more primers. In various embodiments, the library comprises at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different primers. In various embodiments, the library comprises at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different primer pairs, each primer pair comprising a forward test primer and a reverse test primer, each test primer pair hybridizing to a target locus. In some embodiments, the primer library includes at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different individual primers that each hybridize to a different target locus, and the individual primers are not part of a primer pair. In some embodiments, the library includes both (i) primer pairs and (ii) individual primers that are not part of a primer pair (e.g., universal primers).

Hybrid capture probe refers to any nucleic acid sequence, optionally modified, generated by various methods such as PCR or direct synthesis, that is intended to be complementary to one strand of a specific target DNA sequence in a sample. Exogenous hybrid capture probes can be added to a prepared sample and hybridized through a denaturation-reannealing process to form exogenous fragment-endogenous fragment duplexes. These duplexes can then be physically separated from the sample by various means.

A sequence read refers to data that represents a sequence of nucleotide bases measured using clonal sequencing. Clonal sequencing can produce sequence data that represents a single original DNA molecule or a clone of an original DNA molecule or a cluster of original DNA molecules. A sequence read may also have an associated quality score at each base position of the sequence that indicates the probability that the nucleotide was called correctly.

Mapping a sequence read is the process of determining the starting location of a sequence read in the genome sequence of a particular organism. The starting location of a sequence read is based on the nucleotide sequence similarity between the read and the genome sequence.

Matching copy errors, also known as "matching chromosome aneuploidy" (MCA), refer to a state of aneuploidy in which one cell contains two identical or nearly identical chromosomes. This type of aneuploidy can occur during gamete formation in meiosis and can be referred to as a meiotic nondisjunction error. This type of error can occur in mitosis. Matching trisomy can refer to when three copies of a given chromosome are present in an individual and two of the copies are identical.

Mismatched copy errors, also known as "unique chromosome aneuploidy" (UCA), refer to a state of aneuploidy in which one cell contains two chromosomes that are from the same parent and may be homologous but not identical. This type of aneuploidy may arise during meiosis and may be referred to as a meiotic error. Mismatched trisomy may refer to the case where three copies of a given chromosome are present in an individual, where two of the copies are from the same parent and are homologous but not identical. Note that mismatched trisomy may refer to the case where two homologous chromosomes from one parent are present, where some segments of the chromosome are identical while other segments are merely homologous.

Homologous chromosomes refer to chromosome copies that contain the same set of genes that normally pair up during meiosis.

Identical chromosomes refer to chromosome copies that contain the same set of genes, and for each gene, identical chromosomes have the same set of alleles that are identical or nearly identical.

Allelic dropout (ADO) refers to the situation in which at least one of the base pairs in a set of base pairs from the homologous chromosome at a given allele is not detected.

Locus dropout (LDO) refers to the situation where both base pairs within a set of base pairs from homologous chromosomes at a given allele are not detected.

Homozygosity refers to having similar alleles at corresponding chromosomal loci.

Heterozygosity refers to the possession of nonidentical alleles at corresponding chromosomal loci.

Heterozygosity rate refers to the rate at which individuals in a population have heterozygous alleles at a given locus. Heterozygosity rate can also refer to the expected or measured allele ratio at a given locus in an individual or sample of DNA.

Highly informative single nucleotide polymorphisms (HISNPs) are SNPs in which the fetus has an allele that is not present in the mother's genotype.

A chromosomal region refers to a segment of a chromosome or a complete chromosome.

A chromosomal segment refers to a section of a chromosome that can range in size from a single base pair to an entire chromosome.

Chromosome refers to either a complete chromosome or a segment or section of a chromosome.

Copy refers to the number of copies of a chromosomal segment. It may refer to identical copies of a chromosomal segment, or non-identical homologous copies, where different copies of a chromosomal segment contain a substantially similar set of loci and differ in one or more of the alleles. Note that in some cases of aneuploidy, such as M2 copy error, it is possible to have some identical copies of a given chromosomal segment as well as some non-identical copies of the same chromosomal segment.

A haplotype generally refers to a combination of alleles at multiple loci that are inherited together on the same chromosome. A haplotype can refer to as few as two loci or an entire chromosome, depending on the number of recombination events that have occurred between a given set of loci. A haplotype can also refer to a set of single nucleotide polymorphisms (SNPs) on a single chromatid that are statistically associated.

Haplotype data, also "phased data" or "ordered genetic data", refers to data from a single chromosome of a diploid or polyploid genome, i.e., from either the separated maternal or paternal copy of a chromosome of a diploid genome.

Phasing refers to the act of determining the haplotype genetic data of an individual given unordered, diploid (or polyploid) genetic data. Phasing can refer to the act of determining, for a set of alleles found on a chromosome, which of the two genes of the allele are associated with each of the two homologous chromosomes of the individual.

Phase-specified data refers to genetic data in which one or more haplotypes have been determined.

A hypothesis refers to a set of possible ploidy states at a given set of chromosomes or possible allelic states at a given set of loci. A set of possibilities may contain one or more elements.

Copy number hypothesis, also "ploidy state hypothesis", refers to a hypothesis regarding the number of copies of a chromosome in an individual. It may also refer to a hypothesis regarding the identity of each of the chromosomes, including the parent of origin of each chromosome and which of the two parental chromosomes are present in the individual. It may also refer to a hypothesis regarding which, if any, chromosomes or chromosome segments from related individuals correspond genetically to a given chromosome from an individual.

Target individual refers to an individual for whom a genetic state is to be determined. In some embodiments, only a limited amount of DNA is available from the target individual. In some embodiments, the target individual is a fetus. In some embodiments, there may be more than one target individual. In some embodiments, each fetus born to a pair of parents may be considered a target individual. In some embodiments, the genetic data to be determined is an allele call or a collection of allele calls. In some embodiments, the genetic data to be determined is a ploidy call.

Related individual refers to any individual that is genetically related to the target individual and therefore shares a haplotype block with the target individual. In some situations, a related individual may be a genetic parent of the target individual, or any genetic material derived from a parent, such as a sperm, polar body, embryo, fetus, or child. Related individuals may also refer to siblings, parents, or grandparents.

Sibling refers to any individual who has the same genetic parent as the individual in question. In some embodiments, sibling may refer to a born child, embryo, or fetus, or one or more cells derived from a born child, embryo, or fetus. Sibling may also refer to a haploid individual, e.g., sperm, polar body, or any other collection of haplotype genetic material originating from one parent. An individual may consider itself a sibling.

Fetal refers to "of the fetus" or "of the region of the placenta that is genetically similar to the fetus." In pregnant women, parts of the placenta are genetically similar to the fetus, and free-floating fetal DNA found in maternal blood may originate from parts of the placenta that are genotypically similar to the fetus. Note that the genetic information of half of the fetal chromosomes is inherited from the fetus's mother. In some embodiments, DNA from these maternally inherited chromosomes that originate from fetal cells is considered to be "of fetal origin" rather than "of maternal origin."

DNA of fetal origin refers to DNA that was originally part of a cell whose genotype was essentially equivalent to that of the fetus.

DNA of maternal origin refers to DNA that was originally part of a cell whose genotype was essentially identical to that of the mother.

Offspring may refer to an embryo, blastomere, or fetus. Note that in the presently disclosed embodiments, the concepts described apply equally well to an individual that is a born child, fetus, embryo, or collection of cells derived therefrom. The use of the term offspring is simply meant to connote that the individual referred to as offspring is the genetic descendant of the parent.

Parent refers to an individual's genetic mother or father. An individual generally has two parents, a mother and a father, but this is not necessarily the case, for example in genetic or chromosomal chimerism. A parent can be considered an individual.

Parental context refers to the genetic state of a given SNP on each of the two relevant chromosomes for one or both of the two target parents.

Desired development, as well as "normal development," refers to implantation of a viable embryo in the uterus resulting in pregnancy, and/or continuation of the pregnancy resulting in live birth, and/or the absence of chromosomal abnormalities in the offspring, and/or the absence of other undesirable genetic conditions in the offspring, such as disease-linked genes. The term "desired development" is meant to encompass anything that may be desired by a parent or health care provider. In some cases, "desired development" may refer to a non-viable or viable embryo that is useful for medical research or other purposes.

Uterine insertion refers to the process of transferring an embryo into the uterine cavity in the context of in vitro fertilization.

Maternal plasma refers to the plasma portion of blood from a pregnant woman.

A clinical decision refers to any decision to take or not take an action that has an outcome that affects the health or survival of an individual. In the context of prenatal diagnosis, a clinical decision may refer to the decision to abort or not abort a fetus. A clinical decision may also refer to the decision to perform further testing, to take action to reduce an undesirable phenotype, or to take action to prepare for the birth of a child with an abnormality.

Diagnostic Box refers to a machine or combination of machines designed to perform one or more aspects of the methods disclosed herein. In some embodiments, the Diagnostic Box can be located at a patient care location. In some embodiments, the Diagnostic Box can perform targeted amplification followed by sequencing. In some embodiments, the Diagnostic Box can function alone or with the assistance of a technician.

Informatics-based methods refer to methods that rely heavily on statistics to make sense of large amounts of data. In the context of prenatal diagnosis, informatics-based methods refer to methods designed to determine the ploidy state at one or more chromosomes or the allelic state at one or more alleles not by direct physical measurement of the state, but by statistically inferring the most likely state given large amounts of genetic data, e.g., from molecular arrays or sequencing. In some embodiments of the present disclosure, the informatics-based technique may be one disclosed in this patent. In some embodiments of the present disclosure, the informatics-based technique may be PARENTAL SUPPORT™.

Primary genetic data refers to the analog intensity signal output from a genotyping platform. In the context of SNP arrays, primary genetic data refers to the intensity signal before any genotype calls are made. In the context of sequencing, primary genetic data refers to the analog measurements, similar to a chromatogram, that come out of the sequencer before any base pair identities are determined and before the sequence is mapped to the genome.

Secondary genetic data refers to the processed genetic data output from a genotyping platform. In the context of SNP arrays, secondary genetic data refers to the allele calls made by software associated with the SNP array reader that calls the presence or absence of a given allele in a sample. In the context of sequencing, secondary genetic data refers to the base pair identity of a sequence being determined, and in some cases, the sequence being mapped to a genome as well.

Non-invasive prenatal diagnosis (NPD) or, similarly, "non-invasive prenatal screening" (NPS) refers to a method of determining the genetic status of a fetus during pregnancy by a mother using genetic material found in the mother's blood, which is obtained by drawing the mother's intravenous blood.

Preferential enrichment of DNA corresponding to a locus or preferential enrichment of DNA at a locus refers to any method that results in a higher percentage of DNA molecules in the enriched DNA mixture that correspond to the locus than the percentage of DNA molecules in the pre-enriched DNA mixture that correspond to the locus. The method may include selective amplification of DNA molecules that correspond to the locus. The method may include removing DNA molecules that do not correspond to the locus. The method may include a combination of methods. The degree of enrichment is defined as the percentage of DNA molecules in the enriched mixture that correspond to the locus divided by the percentage of DNA molecules in the pre-enriched mixture that correspond to the locus. Preferential enrichment can be performed at multiple loci. In some embodiments of the present disclosure, the degree of enrichment is greater than 20. In some embodiments of the present disclosure, the degree of enrichment is greater than 200. In some embodiments of the present disclosure, the degree of enrichment is greater than 2,000. When preferential enrichment is performed at multiple loci, the degree of enrichment can refer to the average degree of enrichment of all loci in the set of loci.

Amplification refers to a method for increasing the number of copies of a DNA molecule.

Selective amplification can refer to a method of increasing the number of copies of a particular DNA molecule or a DNA molecule that corresponds to a particular region of DNA. Selective amplification can also refer to a method of increasing the number of copies of a particular target DNA molecule or region of target DNA over increasing unlabeled molecules or regions of DNA. Selective amplification can be a method of preferential enrichment.

A universal priming sequence refers to a DNA sequence that can be added to a population of target DNA molecules, for example, by ligation, PCR, or ligation-mediated PCR. Once added to a population of target molecules, a primer specific to the universal priming sequence can be used to amplify the target population using a single pair of amplification primers. The universal priming sequence is generally not related to the target sequence.

Universal Adapters or "Ligation Adapters" or "Library Tags"
is a DNA molecule that contains a universal priming sequence that can be covalently linked to the 5' and 3' ends of a population of target double-stranded DNA molecules. The addition of adapters provides universal priming sequences at the 5' and 3' ends of the target population from which PCR amplification can be performed, amplifying all molecules from the target population using a single pair of amplification primers.

Targeting refers to methods used to selectively amplify or otherwise preferentially enrich DNA molecules that correspond to a set of loci in a mixture of DNA.

A joint distribution model refers to a model that defines the probability of a defined event with respect to multiple random variables, considering multiple random variables defined over the same probability space, where the probabilities of the variables are related. In some embodiments, degenerate cases can be used, where the probabilities of the variables are not related.

The presently disclosed embodiments are further described with reference to the accompanying drawings, in which like structures are referenced by like numerals throughout the several views. The drawings shown are not necessarily to scale, with emphasis generally being placed upon illustrating the principles of the presently disclosed embodiments.

FIG. 2 is a schematic diagram of the direct multiplex mini-PCR method. FIG. 2 is an illustration of the semi-nested mini-PCR method. FIG. 2 is a diagram illustrating the fully nested mini-PCR method. FIG. 2 is an illustration of the hemi-nested mini-PCR method. FIG. 2 is an illustration of triple hemi-nested mini-PCR. FIG. 2 is a diagram illustrating one-sided nested mini-PCR. FIG. 2 is an explanatory diagram of one-sided mini-PCR method. FIG. 2 is a schematic diagram of the reverse semi-nested mini-PCR method. FIG. 1 is a diagram of some possible workflows for the semi-nested method. FIG. 1 is an illustration of a loop ligation adaptor. FIG. 1 is an illustration of internally tagged primers. 1 is an example of some primers with internal tags. FIG. 1 is an illustration of a method using a primer with a ligation adaptor binding region. 1 is a graph showing simulated ploidy calling accuracy for counting methods using two different analytical techniques. FIG. 1 shows the ratio of the two alleles for multiple SNPs in cell lines from experiment 4. FIG. 1 shows the ratio of the two alleles for multiple SNPs in cell lines from experiment 4, separated by chromosome. 17A-D show the ratios of the two alleles for multiple SNPs in plasma samples of four pregnant women, separated by chromosome. 17A-D show the ratios of the two alleles for multiple SNPs in plasma samples of four pregnant women, separated by chromosome. 17A-D show the ratios of the two alleles for multiple SNPs in plasma samples of four pregnant women, separated by chromosome. 17A-D show the ratios of the two alleles for multiple SNPs in plasma samples of four pregnant women, separated by chromosome. 1 is a graph showing the proportion of data that can be explained by binomial variance before and after data correction. 1 is a graph showing the relative enrichment of fetal DNA in samples following a short library preparation protocol. 1 is a graph comparing direct PCR and semi-nested methods in terms of read depth. Graph showing comparison of direct PCR read depth for three genomic samples. Graph showing comparison of semi-nested mini-PCR read depth for three samples. 1 is a graph showing a comparison of read depth for 1,200-plex and 9,600-plex reactions. Graph showing read number ratios in three chromosomes for six types of cells. FIG. 1 shows allelic ratios for three chromosomes for two three-cell reactions and for a third reaction performed on 1 ng of genomic DNA. FIG. 1 shows allelic ratios for two single-cell reactions in three chromosomes. FIG. 1 shows the number of loci with a particular minor allele frequency targeted by each of the two primer libraries. FIG. 28A: Electrophoresis graph of PCR products. 28B-28M are electropherograms of lanes 1-12, respectively, in FIG. 28A. 28B-28M are electropherograms of lanes 1-12, respectively, in FIG. 28A. 28B-28M are electropherograms of lanes 1-12, respectively, in FIG. 28A. Figures 29A-29E: Image representation of the method of the present invention for determining fetal aneuploidy (Figure 29A). Maternal and paternal genotype data (from blood or cheek swabs) and crossover frequency data from the HapMap database are used to generate multiple independent hypotheses for each possible fetal ploidy state in silico (Figure 29B). Each of these hypotheses is expanded to include sub-hypotheses that take into account different possible crossover points. The data model predicts the likely sequencing data (predicted allele distribution) for each hypothetical fetal genotype and different fetal cfDNA percentages and compares them to the actual sequencing data (Figure 29C). The likelihood for each hypothesis is determined using Bayesian statistics. In this hypothetical example, the hypothesis with the highest likelihood (euploidy) is determined (Figure 29D). The individual likelihoods of each copy number hypothesis family (monosomy, disomy, or triploidy) in Figure 29C are summed. The maximum likelihood hypothesis was called as the ploidy state, accounting for the fetal fraction and showing sample-specific calculation accuracy (FIG. 29E). 30A-30H: Representative graphical representations of euploidy (FIGS. 30A-30C), monosomy (FIG. 30D), and trisomy (FIGS. 30E-30H). In all plots, the x-axis represents the linear position of the individual polymorphic locus along the respective chromosome (shown at the bottom of the plot), and the y-axis represents the number of A allele reads as a percentage of the total (A+B) allele reads. Maternal and fetal genotypes, as well as the y-axis location of the band centers, are shown to the right of the plots. If needed for ease of visualization, plots can be color-coded according to maternal genotype, with red indicating AA maternal genotype, blue indicating BB maternal genotype, and green indicating AB maternal genotype. If needed, maternal allele contributions can be indicated by color in the "fetal genotype" column. Allele contributions are expressed in the format maternal|fetal, such as AA|AB for AA maternal and AB fetal alleles. FIG. 30A shows a plot generated when there are two chromosomes and the fetal cfDNA fraction is 0%. This plot is from a non-pregnant woman and therefore represents a pattern when the genotype is entirely maternal. Thus, the allele clusters are centered around 1 (AA alleles), 0.5 (AB alleles), and 0 (BB alleles). FIG. 30B shows a plot generated when there are two chromosomes and the fetal fraction is 12%. Due to the contribution of fetal alleles to the proportion of A allele reads, the position of some allele spots shifts up and down along the y-axis. Therefore, the bands are centered around 1 (AA|AA alleles), 0.94 (AA|AB alleles), 0.56 (AB|AA alleles), 0.50 (AB|AB alleles), 0.44 (AB|BB alleles), 0.06 (BB|AB alleles), and 0 (BB|BB alleles). FIG. 30C shows a plot generated when two chromosomes are present and the fetal fraction is 26%. The pattern is easily visible, including two red and two blue peripheral bands and a triplet of central green bands (colors not shown). The bands are centered at 1 (AA|AA alleles), 0.87 (AA|AB alleles), 0.63 (AB|AA alleles), 0.50 (AB|AB alleles), 0.37 (AB|BB alleles), 0.13 (BB|AB alleles), and 0 (BB|BB alleles). FIG. 30D shows a plot generated when one chromosome is present and the fetal fraction is 26%. The hallmark pattern of one outer red and one outer blue peripheral band, and two central green bands, indicates maternally inherited monosomy (colors not shown). Because the fetus contributes only a single allele (A or B) to the allelic reads, the inner peripheral red and blue bands are absent and the central triplet band is compressed to two bands (colors not shown in figure). The bands are centered at 1 (AA|A allele), 0.57 (AB|A allele), 0.43 (AB|B allele), and 0 (BB|B allele). Figure 30E is a plot generated when three chromosomes are present and the fetal fraction is 27%. This pattern of two red and two blue peripheral bands and two central green bands indicates a maternally inherited meiotic trisomy (colors not shown in figure). The bands are centered at 1 (AA|AAA alleles), 0.88 (AA|AAB alleles), 0.56 (AB|AAB alleles), 0.44 (AB|ABB alleles), 0.12 (BB|ABB alleles), and 0 (BB|BBB alleles). Figure 30F is a plot generated when three chromosomes are present and the fetal fraction is 14%. This pattern of three red and three blue peripheral bands and two central green bands indicates a paternally inherited meiotic trisomy (colors not shown in figure). The bands are centered at 1 (AA|AAA alleles), 0.93 (AA|AAB alleles), 0.87 (AA|ABB alleles), 0.60 (AB|AAA alleles), 0.53 (AB|AAB alleles), 0.47 (AB|ABB alleles), 0.40 (AB|BBB alleles), 0.13 (BB|AAB alleles), 0.07 (BB|ABB alleles), and 0 (BB|BBB alleles). Figure 30G is a plot generated when three chromosomes are present and the fetal fraction is 35%. This pattern of two red and two blue peripheral bands and four green bands indicates a maternally inherited mitotic trisomy (colors not shown in figure). The bands are centered at 1 (AA|AAA alleles), 0.85 (AA|AAB alleles), 0.72 (AB|AAA alleles), 0.57 (AB|AAB alleles), 0.43 (AB|ABB alleles), 0.28 (AB|BBB alleles), 0.15 (BB|ABB alleles), and 0 (BB|BBB alleles). Figure 30H is a plot generated when three chromosomes are present and the fetal fraction is 25%. This pattern of two red and two blue peripheral bands and four central green bands indicates a paternally inherited mitotic trisomy (colors not shown in the figure). This pattern can be distinguished from that of a maternally inherited mitotic trisomy (as in Figure 30G) by the location of the inner peripheral bands. Specifically, the bands are centered at 1 (AA|AAA alleles), 0.78 (AA|ABB alleles), 0.67 (AB|AAA alleles), 0.56 (AB|AAB alleles), 0.44 (AB|ABB alleles), 0.33 (AB|BBB alleles), 0.22 (BB|AAB alleles), and 0 (BB|BBB alleles). 30A-30H: Representative graphical representations of euploidy (FIGS. 30A-30C), monosomy (FIG. 30D), and trisomy (FIGS. 30E-30H). In all plots, the x-axis represents the linear position of the individual polymorphic locus along the respective chromosome (shown at the bottom of the plot), and the y-axis represents the number of A allele reads as a percentage of the total (A+B) allele reads. Maternal and fetal genotypes, as well as the y-axis location of the band centers, are shown to the right of the plots. If needed for ease of visualization, plots can be color-coded according to maternal genotype, with red indicating AA maternal genotype, blue indicating BB maternal genotype, and green indicating AB maternal genotype. If needed, maternal allele contributions can be indicated by color in the "fetal genotype" column. Allele contributions are expressed in the format maternal|fetal, such as AA|AB for AA maternal and AB fetal alleles. FIG. 30A shows a plot generated when there are two chromosomes and the fetal cfDNA fraction is 0%. This plot is from a non-pregnant woman and therefore represents a pattern when the genotype is entirely maternal. Thus, the allele clusters are centered around 1 (AA alleles), 0.5 (AB alleles), and 0 (BB alleles). FIG. 30B shows a plot generated when there are two chromosomes and the fetal fraction is 12%. Due to the contribution of fetal alleles to the proportion of A allele reads, the position of some allele spots shifts up and down along the y-axis. Therefore, the bands are centered around 1 (AA|AA alleles), 0.94 (AA|AB alleles), 0.56 (AB|AA alleles), 0.50 (AB|AB alleles), 0.44 (AB|BB alleles), 0.06 (BB|AB alleles), and 0 (BB|BB alleles). FIG. 30C shows a plot generated when two chromosomes are present and the fetal fraction is 26%. The pattern is easily visible, including two red and two blue peripheral bands and a triplet of central green bands (colors not shown). The bands are centered at 1 (AA|AA alleles), 0.87 (AA|AB alleles), 0.63 (AB|AA alleles), 0.50 (AB|AB alleles), 0.37 (AB|BB alleles), 0.13 (BB|AB alleles), and 0 (BB|BB alleles). FIG. 30D shows a plot generated when one chromosome is present and the fetal fraction is 26%. The hallmark pattern of one outer red and one outer blue peripheral band, and two central green bands, indicates maternally inherited monosomy (colors not shown). Because the fetus contributes only a single allele (A or B) to the allelic reads, the inner peripheral red and blue bands are absent and the central triplet band is compressed to two bands (colors not shown in figure). The bands are centered at 1 (AA|A allele), 0.57 (AB|A allele), 0.43 (AB|B allele), and 0 (BB|B allele). Figure 30E is a plot generated when three chromosomes are present and the fetal fraction is 27%. This pattern of two red and two blue peripheral bands and two central green bands indicates a maternally inherited meiotic trisomy (colors not shown in figure). The bands are centered at 1 (AA|AAA alleles), 0.88 (AA|AAB alleles), 0.56 (AB|AAB alleles), 0.44 (AB|ABB alleles), 0.12 (BB|ABB alleles), and 0 (BB|BBB alleles). Figure 30F is a plot generated when three chromosomes are present and the fetal fraction is 14%. This pattern of three red and three blue peripheral bands and two central green bands indicates a paternally inherited meiotic trisomy (colors not shown in figure). The bands are centered at 1 (AA|AAA alleles), 0.93 (AA|AAB alleles), 0.87 (AA|ABB alleles), 0.60 (AB|AAA alleles), 0.53 (AB|AAB alleles), 0.47 (AB|ABB alleles), 0.40 (AB|BBB alleles), 0.13 (BB|AAB alleles), 0.07 (BB|ABB alleles), and 0 (BB|BBB alleles). Figure 30G is a plot generated when three chromosomes are present and the fetal fraction is 35%. This pattern of two red and two blue peripheral bands and four green bands indicates a maternally inherited mitotic trisomy (colors not shown in figure). The bands are centered at 1 (AA|AAA alleles), 0.85 (AA|AAB alleles), 0.72 (AB|AAA alleles), 0.57 (AB|AAB alleles), 0.43 (AB|ABB alleles), 0.28 (AB|BBB alleles), 0.15 (BB|ABB alleles), and 0 (BB|BBB alleles). Figure 30H is a plot generated when three chromosomes are present and the fetal fraction is 25%. This pattern of two red and two blue peripheral bands and four central green bands indicates a paternally inherited mitotic trisomy (colors not shown in the figure). This pattern can be distinguished from that of a maternally inherited mitotic trisomy (as in Figure 30G) by the location of the inner peripheral bands. Specifically, the bands are centered at 1 (AA|AAA alleles), 0.78 (AA|ABB alleles), 0.67 (AB|AAA alleles), 0.56 (AB|AAB alleles), 0.44 (AB|ABB alleles), 0.33 (AB|BBB alleles), 0.22 (BB|AAB alleles), and 0 (BB|BBB alleles). 30A-30H: Representative graphical representations of euploidy (FIGS. 30A-30C), monosomy (FIG. 30D), and trisomy (FIGS. 30E-30H). In all plots, the x-axis represents the linear position of the individual polymorphic locus along the respective chromosome (shown at the bottom of the plot), and the y-axis represents the number of A allele reads as a percentage of the total (A+B) allele reads. Maternal and fetal genotypes, as well as the y-axis location of the band centers, are shown to the right of the plots. If needed for ease of visualization, plots can be color-coded according to maternal genotype, with red indicating AA maternal genotype, blue indicating BB maternal genotype, and green indicating AB maternal genotype. If needed, maternal allele contributions can be indicated by color in the "fetal genotype" column. Allele contributions are expressed in the format maternal|fetal, such as AA|AB for AA maternal and AB fetal alleles. FIG. 30A shows a plot generated when there are two chromosomes and the fetal cfDNA fraction is 0%. This plot is from a non-pregnant woman and therefore represents a pattern when the genotype is entirely maternal. Thus, the allele clusters are centered around 1 (AA alleles), 0.5 (AB alleles), and 0 (BB alleles). FIG. 30B shows a plot generated when there are two chromosomes and the fetal fraction is 12%. Due to the contribution of fetal alleles to the proportion of A allele reads, the position of some allele spots shifts up and down along the y-axis. Therefore, the bands are centered around 1 (AA|AA alleles), 0.94 (AA|AB alleles), 0.56 (AB|AA alleles), 0.50 (AB|AB alleles), 0.44 (AB|BB alleles), 0.06 (BB|AB alleles), and 0 (BB|BB alleles). FIG. 30C shows a plot generated when two chromosomes are present and the fetal fraction is 26%. The pattern is easily visible, including two red and two blue peripheral bands and a triplet of central green bands (colors not shown). The bands are centered at 1 (AA|AA alleles), 0.87 (AA|AB alleles), 0.63 (AB|AA alleles), 0.50 (AB|AB alleles), 0.37 (AB|BB alleles), 0.13 (BB|AB alleles), and 0 (BB|BB alleles). FIG. 30D shows a plot generated when one chromosome is present and the fetal fraction is 26%. The hallmark pattern of one outer red and one outer blue peripheral band, and two central green bands, indicates maternally inherited monosomy (colors not shown). Because the fetus contributes only a single allele (A or B) to the allelic reads, the inner peripheral red and blue bands are absent and the central triplet band is compressed to two bands (colors not shown in figure). The bands are centered at 1 (AA|A allele), 0.57 (AB|A allele), 0.43 (AB|B allele), and 0 (BB|B allele). Figure 30E is a plot generated when three chromosomes are present and the fetal fraction is 27%. This pattern of two red and two blue peripheral bands and two central green bands indicates a maternally inherited meiotic trisomy (colors not shown in figure). The bands are centered at 1 (AA|AAA alleles), 0.88 (AA|AAB alleles), 0.56 (AB|AAB alleles), 0.44 (AB|ABB alleles), 0.12 (BB|ABB alleles), and 0 (BB|BBB alleles). Figure 30F is a plot generated when three chromosomes are present and the fetal fraction is 14%. This pattern of three red and three blue peripheral bands and two central green bands indicates a paternally inherited meiotic trisomy (colors not shown in figure). The bands are centered at 1 (AA|AAA alleles), 0.93 (AA|AAB alleles), 0.87 (AA|ABB alleles), 0.60 (AB|AAA alleles), 0.53 (AB|AAB alleles), 0.47 (AB|ABB alleles), 0.40 (AB|BBB alleles), 0.13 (BB|AAB alleles), 0.07 (BB|ABB alleles), and 0 (BB|BBB alleles). Figure 30G is a plot generated when three chromosomes are present and the fetal fraction is 35%. This pattern of two red and two blue peripheral bands and four green bands indicates a maternally inherited mitotic trisomy (colors not shown in figure). The bands are centered at 1 (AA|AAA alleles), 0.85 (AA|AAB alleles), 0.72 (AB|AAA alleles), 0.57 (AB|AAB alleles), 0.43 (AB|ABB alleles), 0.28 (AB|BBB alleles), 0.15 (BB|ABB alleles), and 0 (BB|BBB alleles). Figure 30H is a plot generated when three chromosomes are present and the fetal fraction is 25%. This pattern of two red and two blue peripheral bands and four central green bands indicates a paternally inherited mitotic trisomy (colors not shown in the figure). This pattern can be distinguished from that of a maternally inherited mitotic trisomy (as in Figure 30G) by the location of the inner peripheral bands. Specifically, the bands are centered at 1 (AA|AAA alleles), 0.78 (AA|ABB alleles), 0.67 (AB|AAA alleles), 0.56 (AB|AAB alleles), 0.44 (AB|ABB alleles), 0.33 (AB|BBB alleles), 0.22 (BB|AAB alleles), and 0 (BB|BBB alleles). 30A-30H: Representative graphical representations of euploidy (FIGS. 30A-30C), monosomy (FIG. 30D), and trisomy (FIGS. 30E-30H). In all plots, the x-axis represents the linear position of the individual polymorphic locus along the respective chromosome (shown at the bottom of the plot), and the y-axis represents the number of A allele reads as a percentage of the total (A+B) allele reads. Maternal and fetal genotypes, as well as the y-axis location of the band centers, are shown to the right of the plots. If needed for ease of visualization, plots can be color-coded according to maternal genotype, with red indicating AA maternal genotype, blue indicating BB maternal genotype, and green indicating AB maternal genotype. If needed, maternal allele contributions can be indicated by color in the "fetal genotype" column. Allele contributions are expressed in the format maternal|fetal, such as AA|AB for AA maternal and AB fetal alleles. FIG. 30A shows a plot generated when there are two chromosomes and the fetal cfDNA fraction is 0%. This plot is from a non-pregnant woman and therefore represents a pattern when the genotype is entirely maternal. Thus, the allele clusters are centered around 1 (AA alleles), 0.5 (AB alleles), and 0 (BB alleles). FIG. 30B shows a plot generated when there are two chromosomes and the fetal fraction is 12%. Due to the contribution of fetal alleles to the proportion of A allele reads, the position of some allele spots shifts up and down along the y-axis. Therefore, the bands are centered around 1 (AA|AA alleles), 0.94 (AA|AB alleles), 0.56 (AB|AA alleles), 0.50 (AB|AB alleles), 0.44 (AB|BB alleles), 0.06 (BB|AB alleles), and 0 (BB|BB alleles). FIG. 30C shows a plot generated when two chromosomes are present and the fetal fraction is 26%. The pattern is easily visible, including two red and two blue peripheral bands and a triplet of central green bands (colors not shown). The bands are centered at 1 (AA|AA alleles), 0.87 (AA|AB alleles), 0.63 (AB|AA alleles), 0.50 (AB|AB alleles), 0.37 (AB|BB alleles), 0.13 (BB|AB alleles), and 0 (BB|BB alleles). FIG. 30D shows a plot generated when one chromosome is present and the fetal fraction is 26%. The hallmark pattern of one outer red and one outer blue peripheral band, and two central green bands, indicates maternally inherited monosomy (colors not shown). Because the fetus contributes only a single allele (A or B) to the allelic reads, the inner peripheral red and blue bands are absent and the central triplet band is compressed to two bands (colors not shown in figure). The bands are centered at 1 (AA|A allele), 0.57 (AB|A allele), 0.43 (AB|B allele), and 0 (BB|B allele). Figure 30E is a plot generated when three chromosomes are present and the fetal fraction is 27%. This pattern of two red and two blue peripheral bands and two central green bands indicates a maternally inherited meiotic trisomy (colors not shown in figure). The bands are centered at 1 (AA|AAA alleles), 0.88 (AA|AAB alleles), 0.56 (AB|AAB alleles), 0.44 (AB|ABB alleles), 0.12 (BB|ABB alleles), and 0 (BB|BBB alleles). Figure 30F is a plot generated when three chromosomes are present and the fetal fraction is 14%. This pattern of three red and three blue peripheral bands and two central green bands indicates a paternally inherited meiotic trisomy (colors not shown in figure). The bands are centered at 1 (AA|AAA alleles), 0.93 (AA|AAB alleles), 0.87 (AA|ABB alleles), 0.60 (AB|AAA alleles), 0.53 (AB|AAB alleles), 0.47 (AB|ABB alleles), 0.40 (AB|BBB alleles), 0.13 (BB|AAB alleles), 0.07 (BB|ABB alleles), and 0 (BB|BBB alleles). Figure 30G is a plot generated when three chromosomes are present and the fetal fraction is 35%. This pattern of two red and two blue peripheral bands and four green bands indicates a maternally inherited mitotic trisomy (colors not shown in figure). The bands are centered at 1 (AA|AAA alleles), 0.85 (AA|AAB alleles), 0.72 (AB|AAA alleles), 0.57 (AB|AAB alleles), 0.43 (AB|ABB alleles), 0.28 (AB|BBB alleles), 0.15 (BB|ABB alleles), and 0 (BB|BBB alleles). Figure 30H is a plot generated when three chromosomes are present and the fetal fraction is 25%. This pattern of two red and two blue peripheral bands and four central green bands indicates a paternally inherited mitotic trisomy (colors not shown in the figure). This pattern can be distinguished from that of a maternally inherited mitotic trisomy (as in Figure 30G) by the location of the inner peripheral bands. Specifically, the bands are centered at 1 (AA|AAA alleles), 0.78 (AA|ABB alleles), 0.67 (AB|AAA alleles), 0.56 (AB|AAB alleles), 0.44 (AB|ABB alleles), 0.33 (AB|BBB alleles), 0.22 (BB|AAB alleles), and 0 (BB|BBB alleles). Figure 31: Graphical representation of euploid (Figure 31A), T13 (Figure 31B), T18 (Figure 31C), T21 (Figure 31D), 45,X (Figure 31E), and 47,XXY (Figure 31F) test samples. Each chromosome is shown at the top of the plot and fetal and maternal genotypes to the right of the plot, with the x-axis representing the linear position of the SNP along each chromosome and the y-axis showing the number of A allele reads as a percentage of the total reads. Note that the cluster position has been changed based on the fetal fraction as described herein. Each spot represents a single SNP locus. Fetal and maternal genotypes are shown to the right of the plot and chromosome identity is shown at the top of the plot. Figure 31: Graphical representation of euploid (Figure 31A), T13 (Figure 31B), T18 (Figure 31C), T21 (Figure 31D), 45,X (Figure 31E), and 47,XXY (Figure 31F) test samples. Each chromosome is shown at the top of the plot and fetal and maternal genotypes to the right of the plot, with the x-axis representing the linear position of the SNP along each chromosome and the y-axis showing the number of A allele reads as a percentage of the total reads. Note that the cluster position has been changed based on the fetal fraction as described herein. Each spot represents a single SNP locus. Fetal and maternal genotypes are shown to the right of the plot and chromosome identity is shown at the top of the plot. Figure 31: Graphical representation of euploid (Figure 31A), T13 (Figure 31B), T18 (Figure 31C), T21 (Figure 31D), 45,X (Figure 31E), and 47,XXY (Figure 31F) test samples. Each chromosome is shown at the top of the plot and fetal and maternal genotypes to the right of the plot, with the x-axis representing the linear position of the SNP along each chromosome and the y-axis showing the number of A allele reads as a percentage of the total reads. Note that the cluster position has been changed based on the fetal fraction as described herein. Each spot represents a single SNP locus. Fetal and maternal genotypes are shown to the right of the plot and chromosome identity is shown at the top of the plot. Figure 31: Graphical representation of euploid (Figure 31A), T13 (Figure 31B), T18 (Figure 31C), T21 (Figure 31D), 45,X (Figure 31E), and 47,XXY (Figure 31F) test samples. Each chromosome is shown at the top of the plot and fetal and maternal genotypes to the right of the plot, with the x-axis representing the linear position of the SNP along each chromosome and the y-axis showing the number of A allele reads as a percentage of the total reads. Note that the cluster position has been changed based on the fetal fraction as described herein. Each spot represents a single SNP locus. Fetal and maternal genotypes are shown to the right of the plot and chromosome identity is shown at the top of the plot. FIG. 1 is a graph showing that the combined birth prevalence of sex chromosome aneuploidies is higher than that of autosomal aneuploidies.

Although the above figures illustrate embodiments disclosed herein, other embodiments are contemplated as noted in the discussion. This disclosure presents embodiments by way of example, and not by way of limitation. Those skilled in the art can devise numerous other modifications and embodiments which fall within the scope and spirit of the principles of the embodiments disclosed herein.

The present invention is based, in part, on the unexpected discovery that a relatively small number of primers in a primer library are often responsible for the generation of a significant amount of amplified primer dimers formed during a multiplex PCR reaction. Methods have been developed to select the least desirable primers for removal from a candidate primer library. By reducing the amount of primer dimers to a negligible amount (approximately 0.1% of the PCR product), these methods allow the resulting primer library to be used to simultaneously amplify multiple target loci in a single multiplex PCR reaction. Primers hybridize to target loci and amplify the target loci without hybridizing to other primers to form amplified primer dimers, thus increasing the number of different target loci that can be amplified. It has also been found that the use of lower than normal primer concentrations and significantly longer annealing times increases the likelihood that primers will hybridize to target loci, rather than hybridizing to each other to form primer dimers.

During PCR amplification and sequencing of the 19,488 target loci in the genomic sample, 99.4-99.7% of these sequencing reads were mapped to the genome and 99.99% were mapped to the target loci. For a plasma sample with 10 million sequencing reads, typically at least 19,350 (99.3%) of the 19,488 target loci were amplified and sequenced. The ability to simultaneously amplify such a large number of target loci at once significantly reduces the time and amount of DNA required to analyze thousands of target loci. For example, DNA from a single cell is sufficient to simultaneously analyze thousands of target loci, which is important in applications where DNA is low, such as genetic testing of single cells from embryos prior to in vitro fertilization or genetic testing of forensic samples containing small amounts of DNA. Furthermore, the ability to analyze the target loci in one reaction volume (e.g., one vessel or well) without splitting the sample into multiple separate reactions reduces possible variability during the reaction. Furthermore, the method has been developed to use a reference standard to correct for possible amplification biases between different target loci. For example, differences in amplification efficiency between target loci due to factors such as GC content may result in different amounts of PCR product at target loci that should actually be produced in equal amounts. The use of a reference standard similar to the target loci allows such amplification biases to be detected and corrected for during quantification of the target loci.

During sequencing of PCR products, artifacts such as primer dimers are detected, which in turn inhibit the detection of the target amplification products. Due to this limitation, microarrays with hybridization probes are often used for detection, because they are less sensitive to primer dimer interference. The high level of multiplexing currently achievable, with minimal non-target amplification products, allows PCR followed by sequencing to be used as an alternative to microarrays.

The multiplex PCR method of the present invention can be used for a variety of applications, such as genotyping, detection of chromosomal abnormalities (e.g., fetal chromosomal aneuploidies), gene mutation and polymorphism (e.g., single nucleotide polymorphism, SNP) analysis, gene deletion analysis, paternity testing, population genetic variation analysis, forensic analysis, disease predisposition measurement, quantitative analysis of mRNA, and detection and identification of infectious agents (e.g., bacteria, parasites, and viruses). The multiplex PCR method can also be used for non-invasive prenatal genetic testing, such as paternity testing or detection of fetal chromosomal abnormalities.

Representative Primer Design Methods Highly multiplexed PCR can often produce a very high percentage of product DNA resulting from non-productive side reactions such as primer dimer formation. In an embodiment, certain primers that are most likely to cause non-productive side reactions can be removed from the primer library to obtain a primer library that results in a high percentage of amplified DNA that maps to the genome. The step of removing problematic primers, i.e., primers that may stabilize dimers in particular, unexpectedly enabled very high PCR multiplexing levels for subsequent analysis by sequencing. In systems such as sequencing, where performance is significantly reduced by primer dimers and/or other adverse products, multiplexing that is more than 10-fold, more than 50-fold, and more than 100-fold higher than other described multiplexing has been achieved. It should be noted that this is in contrast to probe-based detection methods, e.g., microarray, TAQMAN, PCR, where excess primer dimers do not appreciably affect the results. It should also be noted that the general belief in the art is that multiplexed PCR for sequencing is limited to about 100 assays in the same well. Fluidigm and Rain Dance provide platforms for performing 48 or 1000 PCR assays in parallel reactions on a single sample.

There are several methods for selecting primers for libraries that minimize the amount of non-mapping primer dimers or other adverse primer products. Empirical data shows that a small number of "bad" primers are responsible for a large amount of non-mapping primer dimer side reactions. Removing these "bad" primers can increase the percentage of sequence reads that map to the target locus. One method for identifying "bad" primers is to examine the sequencing data of DNA amplified by targeted amplification, and the most frequently observed primer dimers can be removed to produce a primer library that is significantly less likely to result in by-product DNA that does not map to the genome. There are also publicly available programs that can calculate the binding energy of various primer combinations, and removing the primer combinations with the highest binding energy will similarly produce a primer library that is significantly less likely to result in by-product DNA that does not map to the genome.

In some embodiments for primer selection, an initial candidate primer library is created by designing one or more primers or primer pairs for candidate target loci. A set of candidate target loci (e.g., SNPs) can be selected based on publicly available information on desired parameters of the target loci, such as the frequency of the SNPs in a target population or the heterozygosity rate of the SNPs. In one embodiment, PCR primers can be designed using the Primer3 program (worldwide web at primer3.sourceforge.net; libprimer3 release 2.2.3, incorporated herein by reference in its entirety). Optionally, primers can be designed to anneal within a particular annealing temperature range, to have a particular range of GC content, to fall within a particular size range, to produce target amplicons in a particular size range, and/or to have other parameter characteristics. Starting with multiple primers or primer pairs per candidate target locus increases the likelihood that primers or primer pairs for most or all of the target loci will remain in the library. In one embodiment, the selection criteria may be that at least one primer pair per target locus must remain in the library. In that way, most or all of the target loci will be amplified by using the final primer library. This is desirable for applications such as screening for deletions or duplications at multiple sites in a genome or screening for multiple sequences (e.g., polymorphisms or other mutations) associated with disease or increased risk of disease. If a primer pair from the library produces a target amplicon that overlaps with a target amplicon produced by another primer pair, one of the primer pairs can be removed from the library to prevent interference.

In some embodiments, an "undesirability score" (wherein a higher score indicates the least desirability) is calculated (e.g., computationally) for most or all possible combinations of two primers from a library of candidate primers. In various embodiments, an undesirability score is calculated for at least 80, 90, 95, 98, 99, or 99.5% of the possible combinations of candidate primers in the library. Each undesirability score is based at least in part on the likelihood of dimer formation between the two candidate primers. Optionally, the undesirability score may also be based on one or more other parameters selected from the group consisting of the heterozygosity rate of the target locus, the prevalence of a disease associated with a sequence (e.g., a polymorphism) at the target locus, the disease penetrance associated with a sequence (e.g., a polymorphism) at the target locus, the specificity of the candidate primer for the target locus, the size of the candidate primer, the melting temperature of the target amplicon, the GC content of the target amplicon, the amplification efficiency of the target amplicon, and the size of the target amplicon. When multiple factors are considered, the undesirability score may be calculated based on a weighted average of the various parameters. The parameters can be assigned different weights based on their importance to the particular application for which the primer is to be used. In some embodiments, the primer with the highest undesirability score is removed from the library. If the removed primer is a member of a primer pair that hybridizes to one target locus, the other member of the primer pair can also be removed from the library. The primer and removal process can be repeated as necessary. In some embodiments, the selection method is performed until the undesirability scores of all candidate primer combinations remaining in the library are below a minimum threshold. In some embodiments, the selection method is performed until the number of candidate primers remaining in the library is reduced to a desired number.

In various embodiments, after the undesirability scores are calculated, the candidate primers that are part of the greatest number of combinations of two candidate primers that have undesirability scores above a first minimum threshold are removed from the library. This step ignores interactions below the first minimum threshold because these interactions are less important. If the removed primer is a member of a primer pair that hybridizes to one target locus, the other member of the primer pair can be removed from the library. The process of removing primers can be repeated as necessary. In some embodiments, the selection method is performed until all of the undesirability scores for the candidate primer combinations remaining in the library are below the first minimum threshold. If the number of candidate primers remaining in the library is greater than required, the number of primers can be reduced by lowering the first minimum threshold to a second minimum threshold and repeating the process of removing primers. If the number of candidate primers remaining in the library is less than the required number, the method can continue by repeating the process of removing primers using the original candidate primer library by increasing the first minimum threshold to a larger second minimum threshold, thereby leaving more candidate primers in the library. In some embodiments, the selection method is performed until the undesirability scores of all of the complementary primer combinations remaining in the library are equal to or less than the second minimum threshold, or the number of candidate primers remaining in the library is reduced to the desired number.

If necessary, primer pairs that generate target amplicons that overlap with target amplicons generated by other primer pairs can be split into separate amplification reactions. Multiplex PCR amplification reactions may be preferred for applications in which it is desirable to analyze all candidate target loci (rather than excluding candidate target loci from the analysis due to overlapping target amplicons).

These selection methods minimize the number of candidate primers that need to be removed from the library to achieve the desired reduction in primer dimers. By removing a smaller number of candidate primers from the library, the resulting primer library can be used to amplify more (or all) of the target loci.

Multiplexing a large number of primers places significant constraints on the assays that can be included. Assays that interact unintentionally will result in spurious amplification products. The size constraints of miniPCR may result in further constraints. In an embodiment, it is possible to start with a very large number of potential SNP targets (between about 500 and over 1 million) and attempt to design primers to amplify each SNP. If primers can be designed, it is possible to attempt to identify primer pairs that may form spurious products by evaluating the likelihood of spurious primer duplex formation between all of the possible pairs of primers using published thermodynamic parameters for DNA duplex formation. Primer interactions can be ranked by a score function related to the interaction, and primers with the worst interaction scores can be eliminated until the desired number of primers is met. If SNPs with potential heterozygosity are most useful, it is possible to rank the list of assays in a similar manner and select the assay that best matches heterozygosity. Experiments have verified that primers with high interaction scores are most likely to form primer dimers. In highly multiplexed systems, it is not possible to eliminate all spurious interactions, but it is essential to remove the primers or primer pairs with the highest in silico interaction scores, as they dominate the overall reaction and severely limit amplification from the intended target. This procedure has been implemented to generate multiplexed primer sets up to, and in some cases, more than 10,000 primers. The improvement from this procedure is substantial, allowing amplification of more than 80%, 90%, 95%, 98%, and even 99% of the target product, as determined by sequencing of all PCR products, compared to 10% from reactions that did not remove the worst primers. When combined with the previously described partial semi-nested approach, more than 90% and even 95% of the amplification products can be mapped to the target sequence.

It should be noted that there are other methods for determining which PCR probes are likely to form dimers. In some embodiments, analysis of a pool of DNA amplified using a set of non-optimized primers may be sufficient to determine problematic primers. For example, analysis can be performed using sequencing, and the dimers present in the greatest numbers can be determined to be the ones most likely to form dimers and removed.

This method has a number of potential applications, including SNP genotyping, heterozygosity rate determination, copy number measurement, and other targeted sequencing applications. In some embodiments, the primer design method can be used in combination with the mini-PCR methods described elsewhere in this document. In some embodiments, the primer design method can be used as part of a large-scale multiplex PCR method.

The use of tags on primers can reduce amplification and sequencing of primer dimer products. In some embodiments, the primers include an internal region that forms a loop structure with the tag. In certain embodiments, the primers include a 5' region that is specific to the target locus, an internal region that forms a loop structure that is not specific to the target locus, and a 3' region that is specific to the target locus. In some embodiments, the loop region may be located between two binding regions that are designed to bind to adjacent or adjacent regions of the template DNA. In various embodiments, the length of the 3' region is at least 7 nucleotides. In some embodiments, the length of the 3' region is 7 to 20 nucleotides, e.g., 7 to 15 nucleotides, or 7 to 10 nucleotides. In various embodiments, the primers include a 5' region that is not specific to the target locus (e.g., a tag or universal primer binding site), followed by a region that is specific to the target locus, an internal region that forms a loop structure that is not specific to the target locus, and a 3' region that is specific to the target locus. Using tag-primers, the required target-specific sequence can be shortened to less than 20, 15, 12, or even 10 base pairs. This can be serendipitous with standard primer design when the target sequence is fragmented within the primer binding site, or can be designed into the primer design. Advantages of this method include an increase in the number of assays that can be designed for a particular maximum amplicon length, and shortened sequencing of "uninformative" primer sequences. The method can also be used in combination with internal tagging (see elsewhere in this document).

In some embodiments, the relative amount of non-productive products in multiplex target PCR amplification can be reduced by increasing the annealing temperature. When amplifying libraries with the same tag as the target-specific primers, the annealing temperature can be increased compared to genomic DNA, since the tag contributes to primer binding. In some embodiments, a significantly lower primer concentration than previously reported is used, along with a longer annealing time than reported elsewhere. In some embodiments, the annealing time can be greater than 3 minutes, greater than 5 minutes, greater than 8 minutes, greater than 10 minutes, greater than 15 minutes, greater than 20 minutes, greater than 30 minutes, greater than 60 minutes, greater than 120 minutes, greater than 240 minutes, greater than 480 minutes, and even greater than 960 minutes. In some embodiments, a longer annealing time than previously reported is used, which allows for a lower primer concentration. In various embodiments, a longer extension time than usual is used, for example, greater than 3, 5, 8, 10, or 15 minutes. In some embodiments, primer concentrations are as low as 50 nM, 20 nM, 10 nM, 5 nM, 1 nM, and less than 1 μM. Surprisingly, this results in robust performance for highly multiplexed reactions, such as 1,000-plex, 2,000-plex, 5,000-plex, 10,000-plex, 20,000-plex, 50,000-plex, and even 100,000-plex reactions. In an embodiment, amplification uses 1, 2, 3, 4, or 5 cycles performed with long annealing times, followed by PCR cycles with tagged primers and more than the normal annealing times.

To select target locations, one can start with a pool of candidate primer pair designs, create a thermodynamic model of potentially deleterious interactions between primer pairs, and then use this model to eliminate designs that are incompatible with the other designs in the pool.

After the selection process, the primers remaining in the library can be used in any of the methods of the present invention.

Exemplary Primer Libraries In one aspect, the invention features a library of primers, e.g., primers selected using any of the methods of the invention from a library of candidate primers. In some embodiments, the library includes primers that simultaneously hybridize to (or are capable of simultaneously hybridizing to) or simultaneously amplify (or are capable of simultaneously amplifying) at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci in one reaction volume. In various embodiments, the library comprises primers that simultaneously amplify (or are capable of simultaneously amplifying) between 1,000 and 2,000; 2,000 and 5,000; 5,000 and 7,500; 7,500 and 10,000; 10,000 and 20,000; 20,000 and 25,000; 25,000 and 30,000; 30,000 and 40,000; 40,000 and 50,000; 50,000 and 75,000; 75,000 and 100,000 different target loci in one reaction volume. In various embodiments, the library comprises primers that simultaneously amplify (or are capable of simultaneously amplifying) between 1,000 and 100,000 different target loci in one reaction volume, e.g., between 1,000 and 50,000; between 1,000 and 30,000; between 1,000 and 20,000; between 1,000 and 10,000; between 2,000 and 30,000; between 2,000 and 20,000; between 2,000 and 10,000; between 5,000 and 30,000; between 5,000 and 20,000; or between 5,000 and 10,000 different target loci. In some embodiments, the library includes primers that simultaneously amplify (or are capable of simultaneously amplifying) target loci in one reaction volume, such that less than 60, 40, 30, 20, 10, 5, 4, 3, 2, 1, 0.5, 0.25, 0.1, or 0.5% of the amplified products are primer dimers. In various embodiments, the amount of primer dimer amplified products is between 0.5 and 60%, e.g., between 0.1 and 40%, 0.1 and 20%, 0.25 and 20%, 0.25 and 10%, 0.5 and 20%, 0.5 and 10%, 1 and 20%, or 1 and 10%. In some embodiments, the primers simultaneously amplify (or are capable of simultaneously amplifying) target loci in one reaction volume, such that at least 50, 60, 70, 80, 90, 95, 96, 97, 98, 99, or 99.5% of the amplified products are target amplified products. In various embodiments, the amount of amplification product that is a target amplification product is 50-99.5%, e.g., 60-99%, 70-98%, 80-98%, 90-99.5%, or 95-99.5%. In some embodiments, the primers simultaneously amplify (or are capable of simultaneously amplifying) the target loci in one reaction volume, such that at least 50, 60, 70, 80, 90, 95, 96, 97, 98, 99, or 99.5% of the target loci are amplified. In various embodiments, the amount of the target loci amplified is 50-99.5%, e.g., 60-99%, 70-98%, 80-99%, 90-99.5%, 95-99.9%, or 98-99.99%. In some embodiments, the primer library comprises at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 primer pairs, each pair of primers comprising a forward test primer and a reverse test primer, each test primer pair hybridizing to a target locus. In some embodiments, the primer library comprises at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 individual primers, each hybridizing to a different target locus, each individual primer not being part of a primer pair.

In various embodiments, the concentration of each primer is less than 100, 75, 50, 25, 20, 10, 5, 2, or 1 nM, or less than 500, 100, 10, or 1 uM. In various embodiments, the concentration of each primer is between 1 uM and 100 nM, e.g., between 1 uM and 1 nM, 1 to 75 nM, 2 to 50 nM, or 5 to 50 nM. In various embodiments, the GC content of the primers is between 30 to 80%, e.g., between 40 to 70%, or between 50 to 60%. In some embodiments, the range of the GC content of the primers is less than 30, 20, 10, or 5%. In some embodiments, the range of the GC content of the primers is between 5 to 30%, e.g., between 5 to 20% or between 5 to 10%. In some embodiments, the melting temperature (T _m ) of the test primers is between 40-80° C., e.g., between 50-70° C., 55-65° C., or 57-60.5° C. In some embodiments, the T _m is calculated with the primer3 program (libprimer3 release 2.2.3) using the built-in SantaLucia parameters (worldwide web at primer3.sourceforge.net). In some embodiments, the melting temperature range of the primers is less than 15, 10, 5, 3, or 1° C. In some embodiments, the melting temperature range of the primers is between 1-15° C., e.g., between 1-10° C., 1-5° C., or 1-3° C. In some embodiments, the length of the primers is between 15 and 100 nucleotides, e.g., between 15 and 75 nucleotides, between 15 and 40 nucleotides, between 17 and 35 nucleotides, between 18 and 30 nucleotides, between 20 and 65 nucleotides. In some embodiments, the length of the primers is less than 50, 40, 30, 20, 10, or 5 nucleotides. In some embodiments, the length of the primers is between 5 and 50 nucleotides, e.g., between 5 and 40 nucleotides, between 5 and 20 nucleotides, or between 5 and 10 nucleotides. In some embodiments, the length of the target amplicon is between 50 and 100 nucleotides, e.g., between 60 and 80 nucleotides, or between 60 and 75 nucleotides. In some embodiments, the length of the target amplicon is less than 50, 25, 15, 10, or 5 nucleotides. In some embodiments, the length of the target amplicon is between 5 and 50 nucleotides, e.g., between 5 and 25 nucleotides, between 5 and 15 nucleotides, or between 5 and 10 nucleotides.

These primer libraries can be used in any of the methods of the present invention.

Representative Primer Kits In one aspect, the present invention features a kit (e.g., a kit for amplifying a target locus in a nucleic acid sample) that includes any of the primer libraries of the present invention. In some embodiments, a kit can be formulated that includes a plurality of primers designed to realize the methods described in this disclosure. The primers can be outer forward and reverse primers, inner forward and reverse primers as disclosed herein, or primers designed to have low binding affinity to other primers in the kit as disclosed in the primer design section, or hybrid capture probes or pre-circularization probes as described in the relevant sections, or some combination thereof. In an embodiment, a kit for determining the ploidy state of a target chromosome in a gestating fetus, designed for use in the methods disclosed herein, can be constructed that includes a plurality of inner forward primers, and optionally a plurality of inner reverse primers, and optionally outer forward and outer reverse primers, each of which is designed to hybridize to a region of DNA immediately upstream and/or downstream of one of the target sites (e.g., polymorphic sites) on the target chromosome and, optionally, another chromosome. In some embodiments, the primer kit can be used in combination with the diagnostic boxes described elsewhere in this document. In some embodiments, the kit includes instructions for using the library to amplify target loci.

Exemplary Multiplex PCR Methods In one aspect, the invention features a method of amplifying target loci in a nucleic acid sample, the method including: (i) contacting the nucleic acid sample with a library of primers that simultaneously hybridize to at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci to generate a reaction mixture, and (ii) subjecting the reaction mixture to primer extension reaction conditions (e.g., PCR conditions) to generate an amplification product that includes a target amplicon. In some embodiments, the method also includes determining the presence or absence of at least one target amplicon (e.g., at least 50, 60, 70, 80, 90, 95, 96, 97, 98, 99, or 99.5% of the target amplicon). In some embodiments, the method also includes determining the sequence of at least one target amplicon (e.g., at least 50, 60, 70, 80, 90, 95, 96, 97, 98, 99, or 99.5% of the target amplicons). In some embodiments, at least 50, 60, 70, 80, 90, 95, 96, 97, 98, 99, or 99.5% of the target loci are amplified. In various embodiments, less than 60, 50, 40, 30, 20, 10, 5, 4, 3, 2, 1, 0.5, 0.25, 0.1, or 0.05% of the amplicons are primer dimers.

In an embodiment, the methods disclosed herein use highly efficient highly multiplexed targeted PCR to amplify DNA followed by high-throughput sequencing to determine allele frequency at each target locus. It is novel and non-obvious that more than about 50 or 100 PCR primers can be multiplexed in one reaction volume such that the majority of the resulting sequence reads map to the targeted loci. One technique that allows highly multiplexed targeted PCR to be performed in a highly efficient manner involves the design of primers that are unlikely to hybridize with each other. PCR probes, commonly referred to as primers, are selected by creating a thermodynamic model of potentially deleterious interactions between at least 500, at least 1,000, at least 2,000, at least 5,000, at least 7,500, at least 10,000, at least 20,000, at least 25,000, at least 30,000, at least 40,000, at least 50,000, at least 75,000, or at least 100,000 potential primer pairs, or unintended interactions between primers and sample DNA, and then using this model to eliminate designs that are incompatible with other designs in the pool. Another technique that allows highly multiplexed targeted PCR to be performed in a very efficient manner is to perform targeted PCR using a partial or complete nesting approach. Using one or a combination of these techniques, it is possible to multiplex at least 300, at least 800, at least 1,200, at least 4,000, or at least 10,000 primers in a single pool, with the resulting amplified DNA containing a majority of DNA molecules that map to the targeted locus when sequenced. Using one or a combination of these techniques, it is possible to multiplex a large number of primers in a single pool, with the resulting amplified DNA containing more than 50%, more than 60%, more than 67%, more than 80%, more than 90%, more than 95%, more than 96%, more than 97%, more than 98%, more than 99%, or more than 99.5% of DNA molecules that map to the targeted locus.

In some embodiments, detection of target genetic material can be done in a multiplexed fashion. The number of genetic target sequences that can be run in parallel can range from 1-10, 10-100, 100-1,000, 1,000-10,000, 10,000-100,000, 100,000-1,000,000, or 1,000,000-10,000,000. Previous attempts to multiplex more than 100 primers per pool have resulted in significant problems with undesirable side reactions such as primer dimer formation.

Targeted PCR
In some embodiments, PCR can be used to target specific locations in the genome. In plasma samples, the original DNA is highly fragmented (generally less than 500 bp, with an average length of less than 200 bp). In PCR, both forward and reverse primers anneal to the same fragment to allow amplification. Therefore, if the fragment is short, the PCR assay must amplify a relatively short region as well. If the polymorphic position is too close to the polymerase binding site, as in MIPS, it will bias the amplification from different alleles. Currently, PCR primers that target polymorphic regions, such as those containing SNPs, are generally designed so that the 3' end of the primer hybridizes to the base immediately adjacent to one or more polymorphic bases. In some embodiments of the present disclosure, the 3' end of both the forward and reverse PCR primers are designed to hybridize to a base that is one or a few positions away from the position of the mutation (polymorphic site) of the targeted allele. The number of bases between the polymorphic site (SNP or other type) and the base to which the 3' end of the primer is designed to hybridize can be 1 base, 2 bases, 3 bases, 4 bases, 5 bases, 6 bases, 7-10 bases, 11-15 bases, or 16-20 bases. The forward and reverse primers can be designed to hybridize a different number of bases away from the polymorphic site.

Although a large number of PCR assays can be generated, interactions between different PCR assays make it difficult to multiplex them beyond about 100 assays. Using various complex molecular techniques, the level of multiplexing can be increased, but still limited to less than 100, perhaps 200, or possibly 500 assays per reaction. Samples with large amounts of DNA can be split into multiple sub-reactions and then recombined prior to sequencing. For samples with limited overall samples or some subpopulations of DNA molecules, splitting the sample will introduce statistical noise. In certain embodiments, low or limited amounts of DNA can refer to amounts less than 10 pg, between 10 pg and 100 pg, between 100 pg and 1 ng, between 1 ng and 10 ng, or between 10 ng and 100 ng. It should be noted that this method is particularly useful for small amounts of DNA, where other methods involving splitting into multiple pools can cause significant problems associated with the introduction of stochastic noise, but the method provides the benefit of minimizing bias when performed on samples of any amount of DNA. In these situations, a universal preamplification step can be used to increase the overall sample volume. Ideally, this preamplification step should not appreciably alter the allele distribution.

In an embodiment, the disclosed method allows for the generation of PCR products specific for a large number of target loci, specifically 1,000-5,000 loci, 5,000-10,000 loci, or more than 10,000 loci, from a limited sample, such as a single cell or DNA from a bodily fluid, for genotyping by sequencing or some other genotyping method. Currently, performing multiplex PCR reactions for more than 5 to 10 targets presents a major challenge and is often hindered by primer by-products, e.g., primer dimers, and other artifacts. When detecting target sequences using microarrays with hybridization probes, primer dimers and other artifacts can be ignored, as they are not detected. However, when using sequencing as a method of detection, the majority of the sequencing reads sequence such artifacts and do not sequence the desired target sequences in the sample. Prior art methods used to multiplex greater than 50 or 100 reactions in one reaction volume followed by sequencing typically result in greater than 20%, and often greater than 50%, often greater than 80%, and in some cases greater than 90% off-target sequence reads.

Typically, to perform targeted sequencing on a large number (n) of targets in a sample (>50, >100, >500 or >1,000), the sample can be split into several parallel reactions and one individual target can be amplified. This can be done in PCR multi-well plates or in commercial platforms such as the FLUIDIGM ACCESS ARRAY (48 reactions per sample in a microfluidic chip) or the DROPLET PCR from RAIN DANCE TECHNOLOGY (100-1000 targets). Unfortunately, these split-and-pool methods are problematic for samples with limited amounts of DNA, as there are often insufficient copies of the genome to ensure that there is one copy of each region of the genome in each well. This is a particularly significant problem when targeting polymorphic loci, where the relative proportions of alleles at polymorphic loci are needed because the stochastic noise introduced by splitting and pooling causes a very poorly accurate measurement of the proportion of alleles that were present in the original sample of DNA. Described herein are methods for effective and efficient amplification of many PCR reactions that are applicable when only limited amounts of DNA are available. In certain embodiments, the methods can be applied to analyze single cells, body fluids, mixtures of DNA, such as free-floating DNA found in maternal plasma, biopsies, environmental samples and/or forensic samples.

In some embodiments, targeted sequencing may include one, several, or all of the following steps: a) Generate and amplify libraries with adapter sequences at both ends of DNA fragments; b) After library amplification, split into multiple reactions; c) Generate and optionally amplify libraries with adapter sequences at both ends of DNA fragments; d) Perform 1000-10,000-plex amplification of selected targets using one target-specific "forward" primer and one tag-specific primer per target; e) Perform a second amplification of this product using a "reverse" target-specific primer and one (or more) primers specific for a universal tag that were introduced as part of the target-specific forward primer in the first round; f) Perform 1000-plex preamplification of selected targets for a limited number of cycles; g) Split the product into multiple aliquots and amplify subpools of targets in separate reactions (e.g., 50-500-plex (this method can be used with any plex up to single plex)); h) Pool the products of the parallel subpool reactions. i) During these amplifications, the primers can carry sequencing-compatible tags (partial or full length), allowing the products to be sequenced.

Highly multiplexed PCR
Disclosed herein are methods that allow targeted amplification of over 100 to tens of thousands of target sequences (e.g., SNP loci) from a nucleic acid sample, such as genomic DNA obtained from plasma. The amplified samples are relatively free of primer dimer products and have low allelic bias at the target loci. Analysis of these products can be performed by sequencing if sequencing-compatible adapters are added to the products during or after amplification.

Performing highly multiplexed PCR amplifications using methods known in the art results in an excess of the desired amplification products and the generation of primer dimer products that are not suitable for sequencing. These can be reduced empirically by eliminating the primers that form these products or by performing in silico selection of primers. However, the larger the number of assays, the more challenging this problem becomes.

One solution is to split the 5,000-plex reaction into several lower-plex amplifications, e.g., 100 50-plex reactions or 50 100-plex reactions, or to use microfluidics, or even to split the sample into individual PCR reactions. However, when sample DNA is limited, such as in non-invasive prenatal testing from pregnancy plasma, splitting the sample between multiple reactions should be avoided as this creates bottlenecking.

Described herein is a method for first globally amplifying the plasma DNA of a sample and then splitting the sample into multiple multiplexed target enrichment reactions with a more moderate number of target sequences per reaction. In an embodiment, the disclosed method can be used to preferentially enrich a DNA mixture at multiple loci, the method comprising one or more of the following steps: generating and amplifying a library from a mixture of DNA, where the molecules in the library have adapter sequences ligated to both ends of the DNA fragments, splitting the amplified library into multiple reactions, and performing a first round of multiplex amplification of selected targets using one target-specific "forward" primer per target and one or more adapter-specific universal "reverse" primers. In an embodiment, the disclosed method further comprises performing a second amplification using a "reverse" target-specific primer and one or more primers specific to the universal tag introduced as part of the target-specific forward primer in the first round. In an embodiment, the method may involve a fully nested PCR approach, a hemi-nested PCR approach, a semi-nested PCR approach, a one-sided fully nested PCR approach, a one-sided hemi-nested PCR approach, or a one-sided semi-nested PCR approach. In an embodiment, the disclosed method is used to preferentially enrich a DNA mixture at multiple loci, the method includes performing multiplexed pre-amplification of selected targets with a limited number of cycles, dividing the product into multiple aliquots, amplifying subpools of targets in individual reactions, and pooling the products of parallel subpool reactions. It is noted that this approach can be used to perform targeted amplification with low levels of allelic bias for 50-500 loci, for 500-5,000 loci, for 5,000-50,000 loci, or even for 50,000-500,000 loci. In an embodiment, the primers carry partial or full-length sequencing-compatible tags.

The workflow may involve (1) extracting DNA, such as plasma DNA; (2) preparing a fragment library with universal adaptors at both ends of the fragments; (3) amplifying the library using universal primers specific for the adaptors; (4) dividing the amplified sample "library" into multiple aliquots; (5) performing multiplex (e.g., about 100-plex, 1,000 or 10,000-plex with one target-specific primer and tag-specific primers per target) amplification on the aliquots; (6) pooling the aliquots of one sample; (7) barcoding the samples; (8) mixing the samples and adjusting the concentration; and (9) sequencing the samples. The workflow may include multiple substeps that contain one of the steps listed (e.g., the step of preparing the library in step (2) may involve three enzymatic steps (blunting, dA tailing and adaptor ligation) and three purification steps). Workflow steps can be combined, separated, or performed in a different order (e.g., barcoding and pooling samples).

It is important to note that the library amplification can be performed in a biased manner to amplify short fragments more efficiently. In this way, it is possible to preferentially amplify shorter sequences, e.g. mononucleosomal DNA fragments, as cell-free fetal DNA (of placental origin) found in the circulation of pregnant women. Note that the PCR assay may have a tag, e.g. a sequencing tag (usually a 15-25 base truncated form). After multiplexing, the PCR multiplexed products of the samples are pooled and then the tagging is completed (including barcoding) by tag-specific PCR (which can also be done by ligation). Also, the complete sequencing tag can be added to the same reaction as the multiplexing. In the first cycle, the target can be amplified with target-specific primers, after which the tag-specific primers predominate to complete the complete SQ adapter sequence. The PCR primers do not have to carry tags. The sequencing tag can be added to the amplification product by ligation.

In an embodiment, the amplified material can be evaluated using highly multiplexed PCR followed by clonal sequencing for various applications such as detecting fetal aneuploidy. While conventional multiplexed PCR evaluates up to 50 loci simultaneously, the techniques described herein may be used to simultaneously evaluate more than 50 loci, more than 100 loci, more than 500 loci, more than 1,000 loci, more than 5,000 loci, more than 10,000 loci, more than 50,000 loci, and more than 100,000 loci simultaneously. Experiments have shown that up to, including, and more than 10,000 distinct loci can be simultaneously evaluated in a single reaction with sufficient efficiency and specificity to perform non-invasive prenatal aneuploidy diagnosis and/or copy number calling with high accuracy. The assay can be combined in a single reaction with the entire sample, such as a cfDNA sample isolated from maternal plasma, a fraction thereof, or a further processed derivative of the cfDNA sample. The sample (e.g., cfDNA or derivatives) can also be split into multiple parallel multiplex reactions. Optimal sample splitting and multiplexing is determined by trading off various performance specifications. Because of the limited amount of material, splitting the sample into multiple fractions can result in increased sampling noise, handling time, and potential for error. Conversely, high multiplexing can result in increased amounts of spurious amplification and increased amplification inequality, both of which reduce test performance.

Two crucial relevant considerations in the application of the methods described herein are the limited amount of original sample (e.g., plasma) and the number of original molecules in the material from which allele frequency or other measurements are obtained. If the number of original molecules falls below a certain level, random sampling noise can become significant and affect the accuracy of the test. In general, if measurements are performed on samples containing the equivalent of 500-1000 original molecules per target locus, data of sufficient quality can be obtained to perform non-invasive prenatal aneuploidy diagnosis. There are several ways to increase the number of distinct measurements, e.g., to increase the volume of the sample. Each manipulation applied to the sample also potentially results in a loss of material. It is essential to characterize the losses incurred by the various manipulations and to avoid certain manipulations or improve their yield if necessary, in order to avoid losses that could reduce the performance of the test.

In some embodiments, it is possible to reduce potential losses in subsequent steps by amplifying all or a percentage of the original sample (e.g., a cfDNA sample). Various methods are available to amplify all of the genetic material in a sample, thereby increasing the amount available for downstream procedures. In some embodiments, ligation-mediated PCR (LM-PCR) DNA fragments are amplified by PCR after ligation with either one separate adaptor, two separate adaptors, or many separate adaptors. In some embodiments, multiple displacement amplification (MDA) phi-29 polymerase is used to amplify all DNA isothermally. DOP-PCR and its variants use random priming to amplify the original material DNA. Each method has certain characteristics, such as uniformity of amplification across all representative genomic regions, efficiency of capture and amplification of original DNA, and amplification performance as a function of fragment length.

In an embodiment, LM-PCR can be used with a single heteroduplex adapter with a 3' tyrosine. The heteroduplex adapter allows for the use of a single adapter molecule that can be converted into two separate sequences on the 5' and 3' ends of the original DNA fragment during the first round of PCR. In an embodiment, it is possible to fractionate the amplified library by size separation, or the products such as AMPURE, TASS, or other similar methods. Prior to ligation, the sample DNA is blunt-ended and then a single adenosine base is added to the 3' end. Prior to ligation, the DNA can be cleaved using a restriction enzyme or some other cleavage method. During ligation, the 3' adenosine of the sample fragment and the complementary 3' tyrosine overhang of the adapter can enhance ligation efficiency. The extension step of the PCR amplification may be limited in terms of time to reduce amplification from fragments longer than about 200 bp, about 300 bp, about 400 bp, about 500 bp, or about 1,000 bp. Since the longer DNA found in maternal plasma is almost exclusively maternal, this may result in a 10-50% enrichment of fetal DNA and improved test performance. Several reactions were performed using conditions specified by the commercial kit, resulting in successful ligation of less than 10% of the sample DNA molecules. Sequential optimization of the reaction conditions for this improved ligation to approximately 70%.

Mini-PCR
The Mini-PCR method described below is suitable for samples containing short, digested, or fragmented nucleic acids such as cfDNA. Conventional PCR assay design results in a high loss of characteristic fetal molecules, which can be significantly reduced by designing a very short PCR assay, referred to as a mini-PCR assay. Fetal cfDNA in maternal serum is highly fragmented, with fragment sizes distributed in an approximately Gaussian manner, with a mean of 160 bp, a standard deviation of 15 bp, a minimum size of about 100 bp, and a maximum size of about 220 bp. The distribution of fragment start and end positions for targeted polymorphisms is not necessarily random, but varies widely among individual targets and among all targets collectively, and a polymorphic site at one particular target locus may occupy any position from start to end among the various fragments originating from that locus. It is noted that the term mini-PCR may equally well refer to regular PCR without further restrictions or limitations.

During PCR, amplification occurs only from template DNA fragments that contain both forward and reverse primer sites. Because fetal cfDNA fragments are short, the likelihood that both primer sites are present, and thus the likelihood of a fetal fragment of length L containing both forward and reverse primer sites, is the ratio of the length of the amplicon to the length of the fragment. Under ideal conditions, assays in which the amplicon is 45 bp, 50 bp, 55 bp, 60 bp, 65 bp, or 70 bp will successfully amplify from 72%, 69%, 66%, 63%, 59%, or 56% of the available template fragment molecules, respectively. The length of the amplicon is the distance between the 5' ends of the forward and reverse priming sites. Amplified products of shorter lengths than those commonly used by those skilled in the art may result in more efficient measurement of the desired polymorphic locus with only short sequence reads required. In certain embodiments, a substantial fraction of the amplification products should be less than 100 bp, less than 90 bp, less than 80 bp, less than 70 bp, less than 65 bp, less than 60 bp, less than 55 bp, less than 50 bp, or less than 45 bp.

It should be noted that in methods known in the prior art, short assays such as those described herein are not necessary and are usually avoided because they impose significant constraints on primer design due to limited primer length, annealing properties, and distance between forward and reverse primers.

It should also be noted that amplification bias is potentially present if the 3' end of either primer is within approximately 1-6 bases of the polymorphic site. This single base difference in the site where the polymerase first binds can result in preferential amplification of one allele, which can alter the observed allele frequency and reduce performance. All of these constraints make it very difficult to identify primers that successfully amplify specific loci, and in addition, to design a large collection of primers that are compatible with the same multiplex reaction. In one embodiment, the 3' ends of the inner forward and reverse primers are designed to hybridize to a region of DNA that is upstream of the polymorphic site and separated from it by a small number of bases. Ideally, the number of bases can be between 6 and 10 bases, but can equally well be between 4 and 15 bases, 3 and 20 bases, 2 and 30 bases, or 1 and 60 bases, and essentially the same results can be achieved.

Multiplex PCR may involve a single round of PCR in which all targets are amplified, or it may involve one round of PCR followed by one or more rounds of nested PCR or some variation of nested PCR. Nested PCR consists of one or more subsequent rounds of PCR amplification using one or more new primers that bind at least one base pair further inwards than the primer used in the previous round. Nested PCR reduces the number of false amplification targets by amplifying, in subsequent reactions, only those amplification products from the previous reaction that have the correct internal sequence. Reducing false amplification targets improves the number of useful measurements that can be obtained, especially in sequencing. Nested PCR generally involves designing primers completely inwards to the previous primer binding site, which necessarily increases the minimum DNA segment size required for amplification. For samples such as maternal plasma cfDNA, where the DNA is highly fragmented, the larger assay size reduces the number of distinct cfDNA molecules from which measurements can be obtained. In some embodiments, to counteract this effect, a partial nesting approach can be used in which one or both of the second round primers overlap the first binding site extending internally by several bases, providing additional specificity while minimizing expansion of the overall assay size.

In some embodiments, multiplex pools of PCR assays are designed to amplify SNPs or other polymorphic or non-polymorphic loci on one or more chromosomes that are potentially heterozygous, and these assays are used in a single reaction to amplify DNA. The number of PCR assays can be between 50 and 200 PCR assays, between 200 and 1,000 PCR assays, between 1,000 and 5,000 PCR assays, or between 5,000 and 20,000 PCR assays (50-200plex, 200-1,000plex, 1,000-5,000plex, 5,000-20,000plex, and over 20,000plex, respectively). In an embodiment, a multiplex pool of about 10,000 PCR assays (10,000-plex) is designed to amplify potentially heterozygous SNP loci on chromosomes X, Y, 13, 18, and 21, and chromosomes 1 or 2, and these assays are used in a single reaction to amplify cfDNA obtained from material plasma samples, chorionic villus samples, amniocentesis samples, single or small numbers of cells, other body fluids or tissues, cancer or other genetic material. The SNP frequency of each locus can be determined by a clonal method of sequencing the amplified products or some other method. Statistical analysis of allele frequency distribution or ratios of all assays can be used to determine whether the sample contains one or more trisomies of the chromosomes included in the test. In another embodiment, the original cfDNA sample is split into two samples and parallel 5,000-plex assays are performed. In another embodiment, the original cfDNA sample is split into n samples and parallel (approximately 10,000/n)plex assays are performed, where n is between 2 and 12, or between 12 and 24, or between 24 and 48, or between 48 and 96. Data is collected and analyzed similarly to that already described. Note that this method is equally well applicable to detecting translocations, deletions, duplications, and other chromosomal abnormalities.

In some embodiments, tails that have no homology to the target genome can also be added to the 3' or 5' end of any of the primers. These tails facilitate subsequent manipulations, procedures, or measurements. In some embodiments, the tail sequence can be the same for the target-specific forward and reverse primers. In some embodiments, different tails can be used for the target-specific forward and reverse primers. In some embodiments, multiple different tails can be used for different loci or sets of loci. Certain tails can be shared between all loci or between subsets of loci. For example, forward and reverse tails that correspond to the forward and reverse sequences required for any of the current sequencing platforms allow for direct sequencing after amplification. In some embodiments, the tails can be used as a general priming site that can be used to add other useful sequences between all amplified targets. In some embodiments, the inner primers can contain regions designed to hybridize either upstream or downstream of the target locus (e.g., polymorphic locus). In some embodiments, the primers can contain molecular barcodes. In some embodiments, the primers may contain universal priming sequences designed to enable PCR amplification.

In one embodiment, a 10,000-plex PCR assay pool is created in which the forward and reverse primers have tails that correspond to the required forward and reverse sequences required for a high-throughput sequencing instrument, such as the HISEQ, GAIIX, or MYSEQ available from ILLUMINA. Additionally, an additional sequence is included 5' to the sequencing tails that can be used as a priming site for adding a nucleotide barcode sequence to the amplified product in a subsequent PCR, thereby allowing multiplexed sequencing of multiple samples in a single lane of a high-throughput sequencing instrument.

In one embodiment, a 10,000-plex PCR assay pool is created where the reverse primer has a tail corresponding to the required reverse sequence required for a high-throughput sequencing instrument. After amplification in the first 10,000-plex assay, a subsequent PCR amplification can be performed using a partially nested forward primer (e.g., 6-base nested) for all targets and another 10,000-plex pool with a reverse primer corresponding to the reverse sequencing tail included in the first round. This next round of partially nested amplification with only one target-specific primer and a universal primer limits the size required for the assay, lowering sampling noise but significantly reducing the number of spurious amplicons. Sequencing tags can be added to the added ligation adaptors and/or as part of the PCR probe, so that the tags become part of the final amplicons.

The fetal fraction affects the performance of the test. There are several ways to enrich the fetal fraction of DNA found in maternal plasma. The fetal fraction can be increased by the LM-PCR method already discussed above, as well as by targeted removal of long maternal fragments. In an embodiment, prior to the multiplex PCR amplification of the target loci, an additional multiplex PCR reaction can be performed to selectively remove long and large maternal fragments corresponding to the loci targeted in the subsequent multiplex PCR. Additional primers are designed to anneal to sites that are a greater distance from the polymorphism than would be expected to be present among the cell-free fetal DNA fragments. These primers can be used in a one-cycle multiplex PCR reaction prior to the multiplex PCR of the target polymorphic loci. These distal primers are tagged with a molecule or moiety that may allow selective recognition of the tagged piece of DNA. In an embodiment, these molecules of DNA can be covalently modified with a biotin molecule that allows the removal of the newly formed double-stranded DNA containing these primers after one cycle of PCR. The double-stranded DNA formed during that first round is likely to be of maternal origin. Removal of hybrid material can be accomplished by using magnetic streptavidin beads. There are other methods of tagging that may work equally well. In some embodiments, size selection methods can be used to enrich the sample for shorter strands of DNA, such as, for example, DNA less than about 800 bp, less than about 500 bp, or less than about 300 bp. Amplification of the short fragments then proceeds as normal.

The mini-PCR method described in this disclosure allows for highly multiplexed amplification and analysis of hundreds to thousands or even millions of loci from a single sample in a single reaction. At the same time, detection of the amplified DNA can be multiplexed, and by using barcoding PCR, tens to hundreds of samples can be multiplexed in one sequencing lane. This multiplexed detection has been successfully tested up to 49-plex, allowing for a much higher degree of multiplexing. In effect, this allows for genotyping at thousands of SNPs for hundreds of samples in a single sequencing run. For these samples, the method allows for the determination of genotype and heterozygosity rates, and at the same time, copy number, both of which can be used to detect aneuploidy. This method is particularly useful in detecting aneuploidy in a gestating fetus from free-floating DNA found in maternal plasma. This method can be used as part of a method for sexing the fetus and/or predicting fetal paternity. This method can be used as part of a method for mutation load determination. This method can be used on any amount of DNA or RNA, and the target regions can be SNPs, other polymorphic regions, non-polymorphic regions, and combinations thereof.

In some embodiments, ligation-mediated universal PCR amplification of fragmented DNA can be used. Ligation-mediated universal PCR amplification can be used to amplify plasma DNA, which can then be split into multiple parallel reactions. Ligation-mediated universal PCR amplification can also be used to preferentially amplify short fragments, thereby enriching the fetal fraction. In some embodiments, adding tags to the fragments by ligation can allow for detection of shorter fragments, use of shorter target sequence specific portions of the primers, and/or annealing at higher temperatures to reduce non-specific reactions.

The methods described herein can be used for a number of purposes where there is a collection of target DNA mixed with a certain amount of contaminating DNA. In some embodiments, the target DNA and the contaminating DNA may be from genetically related individuals. For example, genetic abnormalities in a fetus (target) can be detected from maternal plasma containing fetal (target) DNA and also maternal (contaminating) DNA, including whole chromosomal abnormalities (e.g., aneuploidies), partial chromosomal abnormalities (e.g., deletions, duplications, inversions, translocations), polynucleotide polymorphisms (e.g., STRs), single nucleotide polymorphisms, and/or other genetic abnormalities or differences. In some embodiments, the target DNA and the contaminating DNA may be from the same individual, but in the case of, for example, cancer, the target DNA and the contaminating DNA differ by one or more mutations (see, e.g., H. Mamon et al. Preferential Amplification of Apoptotic DNA from Plasma: Potential for Enhancing Detection of Minor DNA Alterations in Circulating DNA. Clinical Chemistry 54:9 (2008)). In some embodiments, the DNA can be found in cell culture (apoptotic) supernatants. In some embodiments, apoptosis can be induced in biological samples (e.g., blood) for subsequent library preparation, amplification and/or sequencing. A number of possible workflows and protocols for achieving this goal are presented elsewhere in this disclosure.

In some embodiments, the target DNA may be from a single cell, a DNA sample consisting of less than one copy of the target genome, small amounts of DNA, DNA from mixed sources (e.g., pregnancy plasma: placenta and maternal DNA; cancer patient plasma and tumor: a mixture of healthy and cancer DNA, transplants, etc.), other bodily fluids, cell cultures, culture supernatants, forensic DNA samples, ancient DNA samples (e.g., insects trapped in amber), other DNA samples, and combinations thereof.

In some embodiments, short amplicon sizes can be used. Short amplicon sizes are particularly suitable for fragmented DNA (see, e.g., A. Sikora et al., Detection of increased amounts of cell-free fetal DNA with short PCR amplicons. Clin Chem. January 2010; 56(1):136-8).

Using short amplicon sizes can provide several important benefits. Short amplicon sizes can provide optimized amplification efficiency. Short amplicon sizes generally result in shorter products and therefore less chance of non-specific priming. Shorter products can be clustered more densely on a sequencing flow cell, as the clusters are smaller. Note that the methods described herein can work equally well for longer PCR amplicons. The amplicon length can be increased if necessary, for example, when sequencing a larger sequence range. Experiments using 146-plex targeted amplification with 100-200 bp long assays as the first step of the nested PCR protocol have been performed on single cells and genomic DNA with positive results.

In some embodiments, the methods described herein can be used to amplify and/or detect SNPs, copy number, nucleotide methylation, mRNA levels, expression levels of other types of RNA, other genetic and/or epigenetic features. The mini-PCR methods described herein can be used in conjunction with next generation sequencing, and the methods can be used in conjunction with other downstream methods, such as microarrays, digital PCR counting, real-time PCR, mass spectrometry, etc.

In some embodiments, the mini-PCR amplification methods described herein can be used as part of a method for accurate quantification of minority populations. The methods can be used for absolute quantification using spiked calibrators. The methods can be used for quantification of variants/low abundance alleles by ultra-deep sequencing and can be performed in a highly multiplexed manner. The methods can be used for standard paternity and identity testing of relatives or ancestry in humans, animals, plants or other living organisms. The methods can be used for forensic testing. The methods can be used for rapid genotyping and copy number analysis (CN) on any type of material, e.g., amniotic fluid and CVS, sperm, products of conception (POC). The methods can be used for single cell analysis, such as genotyping on biopsy samples from embryos. The methods can be used for rapid embryo analysis (less than 1 day, 1 day, or 2 days after biopsy) by targeted sequencing using min-PCR.

In some embodiments, the mini-PCR amplification method can be used for tumor analysis: tumor biopsies are often a mixture of healthy and tumor cells. Targeted PCR allows deep sequencing of SNPs and loci without nearby background sequences. The method can be used for copy number and loss of heterozygosity analysis for tumor DNA. The tumor DNA can be present in many different body fluids or tissues of tumor patients. The method can be used for tumor recurrence detection and/or tumor screening. The method can be used for seed quality control testing. The method can be used for breeding or fishing. Note that any of these methods can be used equally well to target non-polymorphic loci for ploidy calling.

Some references describing some of the basic methods underlying this disclosure include: (1) Wang HY, Luo M, Tereshchenko IV, Fricker DM, Cui X, LiJ Y, Hu G, Chu Y, Azaro MA, Lin Y, Shen L, Yang Q, Kambouris ME, Gao R, Shih W, LiH. Genome Res. 2005 Feb;15(2):276-83. Department of Molecular Genetics, Microbiology and Immunology/The Cancer Institute of New Jersey, Robert Wood Johnson Medical School, New Brunswick, New Jersey 08903, USA. (2) High-throughput genotyping of single nucleotide polymorphisms with high sensitivity. Li H, Wang HY, Cui X, Luo M, Hu G, Greenawalt DM, Tereshchenko IV, Li JY, Chu Y, Gao R. Methods Mol Biol. 2007;396-PubMed PMID:18025699. (3) Nested Patch PCR enables highly multiplexed mutation discovery in candidate genes. Varley K. E., Mitra R. D. Genome Res. 2008 Nov;18(11):1844-50. Epub 2008 Oct 10, which describes a method involving multiplexing an average of nine assays for sequencing. Note that the method disclosed herein allows for orders of magnitude more multiplexing than in the above references.

Variants of Targeted PCR - Nesting There are many workflows possible when performing PCR, and several workflows typical of the methods disclosed herein are described. The steps outlined herein are not intended to exclude other possible steps, nor are they meant to imply that any of the steps described herein are necessary for the method to function properly. Numerous parameter variations or other modifications are known in the literature and can be made without affecting the essence of the invention. One particular general workflow is shown below, followed by a number of possible variants. Variants generally refer to possible secondary PCR reactions, e.g., different types of nesting (step 3), that can be performed. It is important to note that variants can be performed at different times or in a different order than explicitly described herein. The examples using polymorphic loci for illustration can be easily adapted to the amplification of non-polymorphic loci, if desired.
1. The DNA in the sample can be appended with ligation adaptors, often referred to as library tags or ligation adaptor tags (LT), which contain universal priming sequences, followed by universal amplification. In an embodiment, this can be done using standard protocols designed to generate sequencing libraries after fragmentation. In an embodiment, the DNA sample can be blunt ended and then an A can be added to the 3' end. A Y-adaptor with a T-overhang can be added and ligated. In some embodiments, other sticky ends other than A or T overhangs can be used. In some embodiments, other adaptors can be added, for example looped ligation adaptors. In some embodiments, the adaptors can have tags designed for PCR amplification.
2. Specific Target Amplification (STA): Hundreds, thousands, tens of thousands, and even hundreds of thousands of targets can be multiplexed with pre-amplification in one reaction volume. STA is typically performed for 10-30 cycles, but can be performed for 5-40 cycles, 2-50 cycles, and even 1-100 cycles. For example, primers can be tailed for a simpler workflow or to avoid sequencing most dimers. Note that dimers of both primers carrying the same tag generally will not be efficiently amplified or sequenced. In some embodiments, between 1 and 10 cycles of PCR can be performed, in some embodiments, between 10 and 20 cycles of PCR can be performed, in some embodiments, between 20 and 30 cycles of PCR can be performed, in some embodiments, between 30 and 40 cycles of PCR can be performed, and in some embodiments, more than 40 cycles of PCR can be performed. The amplification can be linear amplification. The number of PCR cycles can be optimized to result in an optimal depth of read (DOR) profile. Different DOR profiles may be desirable for different purposes. In some embodiments, a more even distribution of reads among all assays is desirable; if the DOR for some assays is very small, the stochastic noise may be too high for the data to be very useful, but if the read depth is very deep, the marginal usefulness of each additional read is relatively small.
Primer tails can improve detection of fragmented DNA from universally tagged libraries. If the library tag and primer tail contain homologous sequences, hybridization can be improved (e.g., lowering the melting temperature (T _M )) and the primer can be extended only if a portion of the primer target sequence is within the DNA fragment of the sample. In some embodiments, 13 or more target specific base pairs can be used. In some embodiments, 10-12 target specific base pairs can be used. In some embodiments, 8-9 target specific base pairs can be used. In some embodiments, 6-7 target specific base pairs can be used. In some embodiments, STA can be performed on pre-amplified DNA, e.g., MDA, RCA, other whole genome amplification or adapter-mediated universal PCR. In some embodiments, STA can be performed on samples in which specific sequences and populations have been enriched or depleted, e.g., by size selection, target capture, directed degradation.
3. In some embodiments, secondary multiplex PCR or primer extension reactions can be performed to increase specificity and reduce unwanted products. For example, full nesting, semi-nesting, hemi-nesting, and/or subdivision of smaller assay pools into parallel reactions are all techniques that can be used to increase specificity. Experiments have shown that splitting a sample into three 400-plex reactions results in product DNA with higher specificity than one 1,200-plex reaction with the exact same primers. Similarly, experiments have shown that splitting a sample into four 2,400-plex reactions results in product DNA with higher specificity than one 9,600-plex reaction with the exact same primers. In some embodiments, target-specific and tag-specific primers of the same and opposite orientations can be used.
4. In some embodiments, it is possible to amplify the DNA sample (diluted, purified or otherwise) produced by the STA reaction using tag-specific primers and "universal amplification", i.e., amplifying many or all of the pre-amplified and tagged targets. The primers may contain additional functional sequences, such as barcodes or complete adapter sequences, required for sequencing in high-throughput sequencing platforms.

These methods can be used to analyze any sample of DNA, and are particularly useful when the sample of DNA is particularly small, or when it is a sample of DNA where the DNA originates from more than one individual, such as maternal plasma. These methods can be used on DNA samples such as single or small numbers of cells, genomic DNA, plasma DNA, amplified plasma libraries, amplified apoptotic supernatant libraries, or other mixed DNA samples. In some embodiments, these methods can be used when cells with different genetic makeup may be present in a single individual, such as cancer or transplants.

Protocol Variations (Variations and/or Additions to the Workflow Above)
Direct multiplex mini-PCR: Specific target amplification (STA) of multiple target sequences using tagged primers is shown in FIG. 1. 101 shows double stranded DNA with polymorphic loci of interest at X. 102 shows double stranded DNA with ligation adaptors added for universal amplification. 103 shows the universally amplified single stranded DNA with hybridized PCR primers. 104 shows the final PCR product. In some embodiments, STA can be performed on more than 100, more than 200, more than 500, more than 1,000, more than 2,000, more than 5,000, more than 10,000, more than 20,000, more than 50,000, more than 100,000, or more than 200,000 targets. In a subsequent reaction, all target sequences are amplified with tag-specific primers and the tags are extended to include all sequences required for sequencing, including sampling indexes. In some embodiments, the primers may not be tagged, or only certain primers may be tagged. Sequencing adaptors may be added by conventional adaptor ligation. In some embodiments, the first primer may carry a tag.

In some embodiments, the primers are designed such that the length of the amplified DNA is unexpectedly short. Prior art has demonstrated that one of skill in the art typically designs an amplification product of 100+ bp. In some embodiments, the amplification product can be designed to be less than 80 bp. In some embodiments, the amplification product can be designed to be less than 70 bp. In some embodiments, the amplification product can be designed to be less than 60 bp. In some embodiments, the amplification product can be designed to be less than 50 bp. In some embodiments, the amplification product can be designed to be less than 45 bp. In some embodiments, the amplification product can be designed to be less than 40 bp. In some embodiments, the amplification product can be designed to be less than 35 bp. In some embodiments, the amplification product can be designed to be between 40 bp and 65 bp.

Experiments were performed using this protocol with 1200-plex amplification. Both genomic DNA and pregnancy plasma were used; approximately 70% of sequence reads were mapped to the target sequence. Details are presented elsewhere in this document. 1042-plex sequencing without assay design and selection resulted in >99% of sequences being primer dimer products.

Sequential PCR: After STA1, multiple aliquots of product can be amplified in parallel using pools of reduced complexity with the same primers. The first amplification can yield enough material to split. This method is particularly good for small samples, e.g., about 6-100 pg, about 100 pg-1 ng, about 1 ng-10 ng, or about 10 ng-100 ng. A 1200-plex to 3x 400-plex protocol was performed. Mapping of sequencing reads increased from about 60-70% in 1200-plex alone to over 95%.

Semi-nested mini-PCR: (see FIG. 2) After STA1, a second STA is performed consisting of a multiplex set of inner nested forward primers (103B, 105b) and one (or a few) tag-specific reverse primers (103A). 101 shows double-stranded DNA with the polymorphic locus of interest at X. 102 shows double-stranded DNA with ligation adaptors added for universal amplification. 103 shows the universally amplified single-stranded DNA with hybridized forward primer B and reverse primer A. 104 shows the PCR product from 103. 105 shows the product from 104 with hybridized nested forward primer b and reverse tag A that is already part of the molecule from the PCR that occurred between 103 and 104. 106 shows the final PCR product. Using this workflow, typically more than 95% of the sequence maps to the intended target. Nested primers may overlap the outer forward primer sequence, but introduce additional 3' terminal bases. In some embodiments, between 1 and 20 extra 3' bases can be used. Experiments have shown that using 9 or more extra 3' bases works well in a 1200-plex design.

Fully nested mini-PCR: (see FIG. 3) After STA step 1, a second multiplex PCR (or reduced complexity parallel m.p. PCR) can be performed with two nested primers carrying tags (A, a, B, b). 101 shows double stranded DNA with polymorphic locus of interest at X. 102 shows double stranded DNA with ligation adaptors added for universal amplification. 103 shows universally amplified single stranded DNA with hybridized forward primer B and reverse primer A. 104 shows PCR product from 103. 105 shows product from 104 with hybridized nested forward primer b and nested reverse primer a. 106 shows final PCR product. In some embodiments, a complete set of two primers can be used. Using experiments using a fully nested mini-PCR protocol, 146-plex amplification was performed on single cells and triplicate cells without step 102, which adds universal ligation adapters and amplifies.

Hemi-nested mini-PCR: (see FIG. 4) It is possible to use target DNA with adaptors at the ends of the fragments. An STA consisting of a multiplex set of forward primers (B) and one (or a few) tag-specific reverse primers (A) is performed. A second STA can be performed using a universal tag-specific forward primer and a target-specific reverse primer. 101 shows double-stranded DNA with a polymorphic locus of interest at X. 102 shows double-stranded DNA with ligation adaptors added for universal amplification. 103 shows universally amplified single-stranded DNA with hybridized reverse primer A. 104 shows PCR product from 103 amplified using reverse primer A and ligation adaptor tag primer LT. 105 shows product from 104 with hybridized forward primer B. 106 shows the final PCR product. In this workflow, target-specific forward and reverse primers are used in separate reactions, which reduces the complexity of the reaction and prevents dimer formation of the forward and reverse primers. Note that in this example, primers A and B can be considered as the first primers, and primers "a" and "b" can be considered as the inner primers. This method is as good as direct PCR, but is a major improvement over direct PCR because it avoids primer dimers. After the first round of the hemi-nested protocol, about 99% of the non-target DNA is typically seen, but after the second round, there is typically a major improvement.

Triple hemi-nested mini-PCR: (see FIG. 5) It is possible to use target DNA with adaptors at the ends of the fragments. An STA consisting of a multiplex set of forward primers (B) and one (or a few) tag-specific reverse primers (A) and (a) is performed. A second STA can be performed using a universal tag-specific forward primer and a target-specific reverse primer. 101 shows double-stranded DNA with a polymorphic locus of interest at X. 102 shows double-stranded DNA with ligation adaptors added for universal amplification. 103 shows universally amplified single-stranded DNA with hybridized reverse primer A. 104 shows PCR product from 103 amplified using reverse primer A and ligation adaptor tag primer LT. 105 shows product from 104 with hybridized forward primer B. 106 shows PCR product from 105 amplified using reverse primer A and forward primer B. 107 shows the product from 106 with the reverse primer "a" hybridized. 108 shows the final PCR product. Note that in this example, primers "a" and B can be considered inner primers and A can be considered the first primer. If desired, both A and B can be considered first primers and "a" can be considered the inner primer. The names of the reverse and forward primers can be switched. In this workflow, target-specific forward and reverse primers are used in separate reactions, which reduces the complexity of the reaction and prevents dimer formation of the forward and reverse primers. This method is a major improvement over direct PCR because it is as good as direct PCR but avoids primer dimers. After the first round of the hemi-nested protocol, typically about 99% non-target DNA is seen, but after the second round, it is typically greatly improved.

One-sided nested mini-PCR: (see FIG. 6) It is possible to use target DNA with adaptors at the ends of the fragments. STA can also be performed using a multiplex set of nested forward primers and a ligation adaptor tag as the reverse primer. A second STA can then be performed using a set of nested forward primers and universal reverse primers. 101 shows double-stranded DNA with a polymorphic locus of interest at X. 102 shows double-stranded DNA with a ligation adaptor added for universal amplification. 103 shows the universally amplified single-stranded DNA with forward primer A hybridized. 104 shows the PCR product from 103 amplified using forward primer A and ligation adaptor tag reverse primer LT. 105 shows the product from 104 with nested forward primer hybridized. 106 shows the final PCR product. In this method, shorter target sequences can be detected by using overlapping primers in the first and second STAs than by standard PCR. The method is generally performed minus a sample of DNA that has already undergone STA step 1 above - universal tagging and amplification, with two nested primers on only one side and a library tag on the other. The method was performed on apoptotic supernatant and pregnancy plasma libraries. Using this workflow, approximately 60% of the sequences were mapped to the intended target. Note that reads that contained reverse adapter sequences were not mapped, so this number is expected to be higher if reads containing reverse adapter sequences were mapped.

One-sided mini-PCR: It is possible to use target DNA with adapters at the ends of the fragments (see Figure 7). STA can be performed with a multiplex set of forward primers and one (or a few) tag-specific reverse primers. 101 shows double-stranded DNA with the polymorphic locus of interest at X. 102 shows double-stranded DNA with ligation adapters added for universal amplification. 103 shows single-stranded DNA hybridized with forward primer A. 104 shows the PCR product from 103 amplified using forward primer A and ligation adapter tag reverse primer LT, which is the final PCR product. This method allows detection of shorter target sequences than by standard PCR. However, since only one target-specific primer is used, it can be relatively non-specific. The efficacy of this protocol is half that of one-sided nested mini-PCR.

Reverse semi-nested mini-PCR: It is possible to use target DNA with adaptors at the ends of the fragments (see FIG. 8). STA can be performed with a multiplex set of forward primers and one (or a few) tag-specific reverse primers. 101 shows double-stranded DNA with a polymorphic locus of interest at X. 102 shows double-stranded DNA with ligation adaptors added for universal amplification. 103 shows single-stranded DNA with hybridized reverse primer B. 104 shows PCR product from 103 amplified using reverse primer B and ligation adaptor tag forward primer LT. 105 shows PCR product 104 with hybridized forward primer A and inner reverse primer "b". 106 shows PCR product amplified from 105 using forward primer A and reverse primer "b", which is the final PCR product. This method allows detection of shorter target sequences than by standard PCR.

There may be further variations that are simply repetitions or combinations of the above methods, for example, double-nested PCR using three sets of primers. Another variation is one-sided semi-nested mini-PCR, where STA can also be performed with multiple sets of nested forward primers and one (or a few) tag-specific reverse primers.

Note that in all of these variations, the identities of the forward and reverse primers can be swapped. Note that in some embodiments, the nested variations can be performed equally well without adding adapter tags and without the initial library preparation that includes a universal amplification step. Note that in some embodiments, additional rounds of PCR can be included with additional forward and/or reverse primers and amplification steps; these additional steps can be particularly useful when it is desirable to further increase the percentage of DNA molecules that correspond to the targeted locus.

Nesting Workflows There are many ways to perform amplification with different degrees of nesting, and different degrees of multiplexing. In Figure 9, a flow chart is shown with some of the possible workflows. Note that the use of 10,000-plex PCR is just an example, and these flow charts work equally well for other degrees of multiplexing.

Loop Ligation Adapters For example, when adding universally tagged adaptors to generate libraries for sequencing, there are a number of ways to ligate the adaptors. One way is to blunt end the sample DNA, perform A-tailing, and ligate with adaptors that have T-overhangs. There are a number of other ways to ligate the adaptors. There are also a number of adaptors that can be ligated. For example, a Y-adapter can be used that consists of two strands of DNA, one strand has a double-stranded region and a region specified by a forward primer region, and the other strand is specified by a double-stranded region complementary to the double-stranded region on the first strand and a region with a reverse primer. When annealed, the double-stranded region may contain a T-overhang to ligate with double-stranded DNA that has an A-overhang.

In an embodiment, the adaptor may be a loop of DNA with complementary terminal regions, containing a forward primer tagged region (LFT), a reverse primer tagged region (LRT), and a cleavage site between the two (see FIG. 10). 101 refers to a double stranded blunt ended target DNA. 102 refers to an A-tailed target DNA. 103 refers to a loop ligation adaptor with a T overhang "T" and a cleavage site "Z". 104 refers to a target DNA with a loop ligation adaptor added. 105 refers to a target DNA with a ligation adaptor added, cleaved at the cleavage site. LFT refers to a ligation adaptor forward tag and LRT refers to a ligation adaptor reverse tag. The complementary regions may end in a T overhang or other feature that can be used to ligate to the target DNA. The cleavage site can be a series of uracils for cleavage along the UNG, or a sequence that can be recognized and cleaved by a restriction enzyme or other cleavage method or simply basic amplification. These adapters can be used, for example, to prepare any library for sequencing. These adapters can be used in combination with any of the other methods described herein, for example, mini-PCR amplification methods.

Internally tagged primers When using sequencing to determine the allele present at a given polymorphic locus, the sequence read typically begins upstream of the primer binding site (a) and then reads the polymorphic site (X). The tags are typically positioned as shown on the left side of FIG. 11. 101 refers to the single stranded target DNA with the polymorphic locus of interest "X" and primer "a" with tag "b" attached. To avoid non-specific hybridization, the primer binding site (the region of the target DNA complementary to "a") is typically 18-30 bp in length. The sequence tag "b" is typically around 20 bp, and although in theory they can be any length longer than about 15 bp, most people use the primer sequences sold by the sequencing platform companies. The distance "d" between "a" and "X" can be at least 2 bp to avoid allelic bias. When performing W multiplex PCR amplification using the methods disclosed herein or other methods, where careful primer design is required to avoid excessive primer-primer interactions, the window of acceptable distance "d" between "a" and "X" can vary considerably: from 2 bp to 10 bp, from 2 bp to 20 bp, from 2 bp to 30 bp, or even from 2 bp to more than 30 bp. Thus, using the primer configuration shown on the left side of FIG. 11, the sequence read must be a minimum of 40 bp to obtain a read long enough to measure the polymorphic locus, and depending on the length of "a" and "d", sequence reads may need to be up to 60 bp or 75 bp. Typically, the longer the sequence read, the more expensive and time it takes to sequence a given number of reads, and therefore minimizing the length of the required reads can save both time and money. Furthermore, since on average, reads of bases earlier in the read are more accurate than those later in the read, reducing the length of the required sequence read can also increase the accuracy of the measurement of the polymorphic region.

In one embodiment, as shown in FIG. 11 at 103, the primer binding site (a), referred to as an internally tagged primer, is divided into multiple segments (a', a'', a''',...) and the sequence tag (b) is placed on the segment of DNA in the middle of the two primer binding sites. This arrangement allows the sequencer to make shorter sequence reads. In one embodiment, a'+a'' should be at least about 18 bp, and may be 30 bp, 40 bp, 50 bp, 60 bp, 80 bp, 100 bp, or more than 100 bp long. In one embodiment, a'' should be at least about 6 bp, and in one embodiment, is between about 8 bp and 16 bp. All other factors being equal, the use of internally tagged primers can trim the length of the required sequence read by at least as much as 6 bp, 8 bp, 10 bp, 12 bp, 15 bp, or even as much as 20 bp or 30 bp. This can result in significant money, time, and accuracy advantages. An example of an internally tagged primer is shown in FIG. 12.

Primers with Ligation Adapter Binding Regions One problem with fragmented DNA is that due to its short length, the polymorphism is more likely to be near the end of the DNA strand than a longer strand (e.g., 101, FIG. 10). Since a primer binding site of appropriate length is required on either side of the polymorphism to capture it by PCR, a significant number of strands of DNA with the target polymorphism will fail to be captured due to insufficient overlap between the primer and target binding sites. In an embodiment, the target DNA 101 can be appended with a ligation adaptor 102, and the target primer 103 can have a region (cr) complementary to the ligation adaptor tag (lt) appended upstream of the designed binding region (a) (see FIG. 13); thus, if the binding region (region of 101 complementary to a) is shorter than the 18 bp typically required for hybridization, the region of the primer complementary to the library tag (cr) can increase the binding energy to the point where PCR can proceed. Note that any specificity lost due to the shorter binding region can be compensated for by other PCR primers with appropriately long target binding regions. Note that this embodiment can be used in combination with direct PCR or any of the other methods described herein, such as nested PCR, semi-nested PCR, hemi-nested PCR, one-sided nested or semi-nested or hemi-nested PCR or other PCR protocols.

When sequencing data is used to determine ploidy in combination with analytical methods that involve comparing observed allele data with predicted allele distributions for various hypotheses, each additional read from an allele with a low read depth provides more information than a read from an allele with a high read depth. Ideally, therefore, one would like to see a uniform read depth (DOR) where each locus has a similar number of representative sequence reads. It is therefore desirable to minimize the variance of the DOR. In an embodiment, it is possible to reduce the coefficient of variation of the DOR (which can be defined as the standard deviation of the DOR/average DOR) by increasing the annealing time. In some embodiments, the annealing temperature may be greater than 2 minutes, greater than 4 minutes, greater than 10 minutes, greater than 30 minutes, and greater than 1 hour or even longer. Since annealing is an equilibrium process, there is no limit to the improvement of the variance of the DOR with increasing annealing time. In an embodiment, the variance of the DOR is reduced by increasing the primer concentration.

Exemplary Whole Genome Amplification Methods In some embodiments, DNA amplification such as whole genome applications can be included to amplify the nucleic acid sample prior to amplifying only the target loci. Amplification of DNA is the process of converting small amounts of genetic material into larger amounts of genetic material containing a collection of similar genetic data, and can be done by a variety of methods including, but not limited to, polymerase chain reaction (PCR). One method of amplifying DNA is whole genome amplification (WGA). There are several methods available for WGA: ligation-mediated PCR (LM-PCR), degenerate oligonucleotide primer PCR (DOP-PCR), and multiple displacement amplification (MDA). In LM-PCR, short DNA sequences called adapters are ligated to the blunt ends of the DNA. These adapters contain universal amplification sequences and are used to amplify the DNA by PCR. In DOP-PCR, random primers, also containing universal amplification sequences, are used in the first round of annealing and PCR. A second round of PCR is then used to further amplify the sequences using the universal primer sequences. MDA uses phi-29 polymerase, a highly processive and non-specific enzyme that replicates DNA and has been used for single cell analysis. The major limitations to the amplification of material from single cells are (1) the need to use extremely dilute DNA concentrations or very small volumes of reaction mixtures, and (2) the difficulty of reliably dissociating DNA from proteins across the entire genome. Nevertheless, single cell whole genome amplification has been used successfully for a variety of applications for many years. There are other methods to amplify DNA from a sample of DNA. DNA amplification converts the initial sample of DNA into a sample of DNA with a similar set of sequences, but in much greater quantities. In some cases, amplification may not be necessary.

In some embodiments, DNA can be amplified using universal amplification, e.g., WGA or MDA. In some embodiments, DNA can be amplified by using targeted amplification, e.g., targeted PCR or circularization probes. In some embodiments, DNA can be preferentially enriched using targeted amplification methods or methods that result in complete or partial separation of desired and undesired DNA, e.g., capture by hybridization techniques. In some embodiments, DNA can be amplified by using a combination of universal amplification and preferential enrichment methods. A fuller description of some of these methods can be found elsewhere in this document.

Exemplary Enrichment and Sequencing Methods In certain embodiments, the methods disclosed herein use selective enrichment techniques that preserve the relative allele frequency present at each target locus (e.g., each polymorphic locus) from a set of target loci (e.g., polymorphic loci) in the original DNA sample. Although enrichment is particularly useful for analyzing polymorphic loci, these enrichment methods can be easily adapted to non-polymorphic loci as needed. In some embodiments, amplification and/or selective enrichment techniques may involve PCR, such as ligation-mediated PCR, fragment capture by hybridization, molecular inversion probes or other circularization probes. In some embodiments, methods for amplification or selective enrichment may involve the use of probes such that when hybridized correctly to the target sequence, the 3' or 5' end of the nucleotide probe is separated from the polymorphic site of the allele by a small number of nucleotides. This separation reduces preferential amplification of one allele, referred to as allelic bias.
This is an improvement over methods involving the use of probes in which the 3' or 5' end of the correctly hybridized probe is directly adjacent to or very close to the polymorphic site of the allele. In an embodiment, probes whose hybridization region may or certainly contains the polymorphic site are excluded. The presence of a polymorphic site at the site of hybridization may cause unequal hybridization at some alleles or may inhibit hybridization altogether, resulting in preferential amplification of certain alleles. These embodiments are an improvement over other methods involving targeted amplification and/or selective enrichment in that they better preserve the original allele frequency at each polymorphic locus of the sample, whether the sample is a pure genomic sample from a single individual or a mixture of individuals.

The use of techniques to enrich and then sequence a sample of DNA at a set of target loci as part of a method for non-invasive prenatal allele or ploidy calling can confer a number of unexpected advantages. In some embodiments of the present disclosure, the method includes measuring genetic data for use in informatics-based methods, such as PARENTAL SUPPORT™ (PS). The end outcome of some of the embodiments is actionable genetic data of the embryo or fetus. There are many methods that can be used to measure genetic data of an individual and/or related individuals as part of the embodied methods. In an embodiment, a method for enriching a set of target alleles is disclosed herein, the method comprising one or more of the following steps: targeted amplification of genetic material, adding a locus-specific oligonucleotide probe, ligating specific DNA strands, isolating the desired set of DNA, removing undesired components of the reaction, detecting specific DNA sequences by hybridization, and detecting the sequence of one or more DNA strands by DNA sequencing methods. In some cases, the DNA strands may refer to the target genetic material, in some cases, the DNA strands may refer to primers, in some cases, the DNA strands may refer to synthesized sequences, or combinations thereof. These steps can be performed in a number of different orders.

For example, a universal amplification step of DNA prior to targeted amplification may confer several advantages, such as elimination of the risk of bottlenecks and reduction of allelic bias. The DNA may be mixed with an oligonucleotide probe capable of hybridizing to two adjacent regions on either side of the target sequence. After hybridization, the ends of the probe may be joined by adding a polymerase, a means of ligation, and any necessary reagents to allow circularization of the probe. After circularization, an exonuclease may be added to digest the non-circularized genetic material, after which the circularized probe may be detected. The DNA may be mixed with a PCR primer capable of hybridizing to two adjacent regions on either side of the target sequence. After hybridization, the ends of the probe may be joined by adding a polymerase, a means of ligation, and any necessary reagents to complete the PCR amplification. The amplified or unamplified DNA may be the target of a hybrid capture probe that targets a set of loci, and after hybridization, the probe may be localized and separated from the mixture to provide a mixture of DNA enriched with the target sequence.

The use of methods to target and subsequently sequence specific loci as part of an allele or ploidy calling methodology can confer a number of unexpected advantages. Some methods by which DNA can be targeted or preferentially enriched include using circularized probes, linked inverted probes (LIPs, MIPs), capture by hybridization methods such as SURESELECT, and targeted PCR or ligation-mediated PCR amplification strategies.

In some embodiments, the disclosed methods include measuring genetic data for use in informatics-based methods, such as PARENTAL SUPPORT™ (PS), further described herein. PARENTAL SUPPORT™ is an informatics-based approach to manipulating genetic data, aspects of which are described herein. The end outcome of some of the embodiments is actionable genetic data of the embryo or fetus, followed by a clinical decision based on the actionable data. The algorithm behind the PS method takes measured genetic data of the target individual, often the embryo or fetus, and measured genetic data from related individuals, and can increase the accuracy with which the genetic status of the target individual is known. In some embodiments, the measured genetic data is used in the context of making ploidy determinations during prenatal genetic diagnosis. In some embodiments, the measured genetic data is used in the context of making ploidy determinations or allele calls on embryos during in vitro fertilization. There are many methods that can be used to measure genetic data of individuals and/or related individuals in the above-mentioned situations. The different methods include a number of steps that often involve amplifying the genetic material, adding oligonucleotide probes, ligating specific DNA strands, isolating the desired set of DNA, removing unwanted components of the reaction, detecting specific DNA sequences by hybridization, detecting the sequence of one or more DNA strands by DNA sequencing methods. In some cases, the DNA strand refers to the target genetic material, in some cases, the DNA strand refers to a primer, in some cases, the DNA strand refers to a synthesized sequence, or a combination thereof. These steps can be performed in a number of different orders.

Note that in theory it is possible to target any number of loci in the genome, anywhere from one locus to over a million loci. When a sample of DNA is targeted and then subjected to sequencing, the percentage of alleles read by the sequencer is enriched relative to the amount they naturally occur in the sample. The degree of enrichment can be anything from 1 percent (or even lower) to 10-fold, 100-fold, 1,000-fold, or even up to a million-fold. There are approximately 3 billion base pairs and nucleotides in the human genome, including approximately 75 million polymorphic loci. The more loci that are targeted, the less enrichment is possible. The fewer the number of loci that are targeted, the greater the degree of enrichment is possible, and the greater the read depth can be achieved at those loci for a given number of sequence reads.

In certain embodiments of the present disclosure, the targeting or preference can be entirely focused on SNPs. In certain embodiments, the targeting or preference can be focused on any polymorphic site. A number of commercial targeting products are available for enriching exons. Surprisingly, targeting exclusively SNPs or exclusively polymorphic loci is particularly advantageous when using methods for NPD that rely on allele distribution. There are also published methods for NPD using sequencing, such as U.S. Pat. No. 7,888,017, which involves a read count analysis where the read count is focused on counting the number of reads that map to a given chromosome, and the analyzed sequence reads are not focused on the region of the genome that is polymorphic. These types of methodologies that do not focus on polymorphic alleles are not as useful as targeting or preferentially enriching a set of alleles.

In certain embodiments of the present disclosure, it is possible to use targeted methods that focus on SNPs to enrich genetic samples in polymorphic regions of the genome. In certain embodiments, it is possible to focus on a small number of SNPs, for example between 1 and 100 SNPs, or a larger number, for example between 100 and 1,000, between 1,000 and 10,000, between 10,000 and 100,000, or more than 100,000 SNPs. In certain embodiments, it is possible to focus on one or a small number of chromosomes that are correlated with live trisomic births, for example chromosomes 13, 18, 21, X and Y, or some combination thereof. In certain embodiments, it is possible to enrich the targeted SNPs by a small factor, for example between 1.01-fold and 100-fold, or by a larger factor, for example between 100-fold and 1,000,000-fold, or more than 1,000,000-fold. In an embodiment of the present disclosure, the targeting method can be used to create a sample of DNA that is preferentially enriched in polymorphic regions of the genome. In an embodiment, the method can be used to create a mixture of DNA with any of these characteristics, where the mixture of DNA contains maternal DNA and also floating fetal DNA. In an embodiment, the method can be used to create a mixture of DNA with any combination of these factors. For example, the method described herein can be used to generate a mixture of DNA that includes maternal DNA and fetal DNA and is preferentially enriched in DNA corresponding to 200 SNPs, all located on either chromosome 18 or chromosome 21, and enriched by an average of 1,000 times. In another example, the method can be used to create a mixture of DNA that is preferentially enriched in 10,000 SNPs, all or most of which are located on chromosomes 13, 18, 21, X and Y, with an average enrichment per locus of more than 500 times. Any of the targeting methods described herein can be used to generate a mixture of DNA that is preferentially enriched at a particular locus.

In some embodiments, the disclosed method further comprises measuring the DNA in the mixed fraction using a high-throughput DNA sequencer, wherein the DNA in the mixed fraction contains a disproportionate number of sequences from one or more chromosomes, the one or more chromosomes being selected from the group including chromosome 13, chromosome 18, chromosome 21, chromosome X, chromosome Y, and combinations thereof.

Three methods are described herein, multiplex PCR, targeted capture by hybridization, and linked inverse probes (LIP), that can be used to obtain and analyze measurements from a sufficient number of polymorphic loci from maternal plasma samples to detect fetal aneuploidy. This is not to exclude other methods of selectively enriching the targeted loci. Other methods can be used equally well without changing the core of the method. In each case, the polymorphisms assayed may include single nucleotide polymorphisms (SNPs), small insertions/deletions, or STRs. The preferred method involves the use of SNPs. Each method generates allele frequency data, and the allele frequency data for each targeted locus and/or the joint allele frequency distribution from these loci can be analyzed to determine fetal ploidy. Each method has its own considerations due to the limited source material and the fact that maternal plasma consists of a mixture of maternal and fetal DNA. This method can be combined with other methods to provide a more accurate determination. In some embodiments, this method can be combined with sequence counting techniques such as those described in U.S. Pat. No. 7,888,017. The described techniques can also be used to detect fetal paternity non-invasively from maternal plasma samples. Additionally, each technique can be applied to other DNA mixtures or pure DNA samples to detect the presence or absence of aneuploid chromosomes, to genotype multiple SNPs from degraded DNA samples, to detect segmental copy number variations (CNVs), to detect other genotypic states of interest, or some combination thereof.

Accurate measurement of allele distribution in a sample Current sequencing techniques can be used to estimate the distribution of alleles in a sample. One such method involves randomly sampling sequences from pooled DNA, called shotgun sequencing. The proportion of specific alleles in sequencing data is generally very low and can be determined by simple statistics. The human genome contains approximately 3 billion base pairs. Therefore, if the sequencing method used produces 100 bp reads, a specific allele is measured approximately once every 30 million sequence reads.

In an embodiment, the disclosed method is used to determine the presence or absence of two or more different haplotypes containing the same set of loci in a sample of DNA from the measured allele distribution of loci from that chromosome. The different haplotypes may represent two different homologous chromosomes from one individual, three different homologous chromosomes from a trisomic individual, three different homologous haplotypes from the mother and fetus, where one of the haplotypes is shared between the mother and fetus, three or four haplotypes from the mother and fetus, where one or two of the haplotypes are shared between the mother and fetus, or other combinations. Alleles that are polymorphic between haplotypes tend to be more informative, but any allele where the mother and father are not both homozygous for the same allele will provide useful information through the measured allele distribution beyond that available from simple read count analysis.

However, shotgun sequencing of such samples is very inefficient, as it results in many sequences for regions that are not polymorphic between the different haplotypes in the sample, or for chromosomes that are not of interest, and therefore does not provide information about the proportion of the target haplotype. Described herein is a method to specifically target and/or preferentially enrich segments of DNA in a sample that are more likely to be polymorphic in the genome, thereby increasing the yield of allelic information obtained by sequencing. It should be noted that for the allelic distribution measured in the enriched sample to truly represent the actual amount present in the target individual, it is crucial that there is little or no preferential enrichment of one allele compared to other alleles at a given locus in the targeted segment. Current methods known in the art for targeting polymorphic alleles are designed to ensure that at least a portion of any alleles present are detected. However, these methods were not designed to measure an unbiased allelic distribution of polymorphic alleles present in the original mixture. It is not self-evident that any particular method of target enrichment can produce an enriched sample in which the measured allele distribution better accurately represents the allele distribution present in the original unamplified sample than any other method. In theory, many enrichment methods can be expected to achieve such a goal, but those skilled in the art are well aware that current amplification, targeting and other preferential enrichment methods have a significant amount of stochastic or deterministic bias. One embodiment of the method described herein allows multiple alleles found in a mixture of DNA corresponding to a given locus in a genome to be amplified or preferentially enriched such that the degree of enrichment of each of the alleles is approximately the same. In other words, the method allows the ratio between the alleles corresponding to each locus to remain essentially the same as in the original DNA mixture, while increasing the relative amount of alleles present in the mixture as a whole. Some reported methods can result in allele biases of more than 1%, more than 2%, more than 5%, or even more than 10%. This preferential enrichment may be due to capture bias when using capture by hybridization techniques, or amplification bias that may be small for each cycle, but may become large when compounded over 20, 30 or 40 cycles. For purposes of this disclosure, the ratio remains essentially the same means that the ratio of alleles in the original mixture divided by the ratio of alleles in the resulting mixture is between 0.95 and 1.05, between 0.98 and 1.02, between 0.99 and 1.01, between 0.995 and 1.005, between 0.998 and 1.002, between 0.999 and 1.001, or between 0.9999 and 1.0001. It should be noted that the calculation of allele ratios presented herein cannot be used to determine the ploidy state of the target individual, but may simply be a metric used to measure allele bias.

In an embodiment, once the mixture is preferentially enriched in a set of target loci, it can be sequenced using any one of the previous, current, or next generation sequencing instruments that sequence clonal samples (samples generated from single molecules; examples include ILLUMINA GAIIx, ILLUMINA HiSeq, Life Technologies SOLiD, 5500XL). The ratio can be assessed by sequencing through specific alleles within the targeted region. These sequencing reads can be analyzed and counted according to the allele type and thus the assignment of the different alleles determined. For variations that are one to a few bases in length, allele detection is performed by sequencing, and it is essential that the sequencing reads span the alleles in question to assess the allelic composition of the captured molecules. The total number of captured molecules assayed for genotype can be increased by increasing the length of the sequencing read. Complete sequencing of all molecules ensures collection of the maximum amount of data available in the enriched pool. However, sequencing is currently expensive, and methods that can measure allele distributions using a small number of sequence reads would be extremely valuable. Furthermore, as read lengths increase, there are technical limits to the maximum length of a possible read as well as limits to accuracy. Alleles of greatest utility are those that are one to a few bases long, but theoretically any allele shorter than the length of the sequencing read can be used. Allelic variation occurs in all forms, but the examples provided herein focus on SNPs or variants contained within only a few adjacent base pairs. Larger variants, e.g., segmental copy number variants, can often be detected by summing up these smaller variations, since the entire population of SNPs within a segment overlaps. Variants larger than a few bases, e.g., STRs, require special consideration and some targeted approach studies, whereas others do not.

There are several targeting approaches that can be used to specifically isolate and enrich one or more variant locations in the genome. Generally, these rely on utilizing non-mutated sequences adjacent to the variant sequence. There have been reports by other researchers related to targeting in sequencing when the substrate is maternal plasma (see, for example, Liao et al., Clin. Chem. 2011; 57(1): pp. 92-101). However, these approaches use targeting probes that target exons and do not focus on targeting polymorphic regions of the genome. In an embodiment, the disclosed method includes using a targeting probe that focuses exclusively or nearly exclusively on the polymorphic region. In an embodiment, the disclosed method includes using a targeting probe that focuses exclusively or nearly exclusively on the SNP. In some embodiments of the present disclosure, the target polymorphic sites consist of at least 10% SNPs, at least 20% SNPs, at least 30% SNPs, at least 40% SNPs, at least 50% SNPs, at least 60% SNPs, at least 70% SNPs, at least 80% SNPs, at least 90% SNPs, at least 95% SNPs, at least 98% SNPs, at least 99% SNPs, at least 99.9% SNPs, or exclusively SNPs.

In an embodiment, the disclosed method can be used to determine genotypes (the base composition of DNA at specific loci) and the relative proportions of these genotypes from a mixture of DNA molecules, which may originate from one or several genetically distinct individuals. In an embodiment, the disclosed method can be used to determine genotypes at a set of polymorphic loci and the relative ratios of the amounts of different alleles present at these loci. In an embodiment, the polymorphic loci may consist entirely of SNPs. In an embodiment, the polymorphic loci may include SNPs, single tandem repeats, and other polymorphisms. In an embodiment, the disclosed method can be used to determine the relative distribution of alleles at a set of polymorphic loci in a mixture of DNA, which includes DNA originating from the mother and DNA originating from the fetus. In an embodiment, the joint allele distribution can be determined for a mixture of DNA isolated from blood from a pregnant woman. In an embodiment, the allele distribution at a set of loci can be used to determine the ploidy state of one or more chromosomes for a pregnant fetus.

In some embodiments, the mixture of DNA molecules may be derived from DNA extracted from multiple cells of one individual. In some embodiments, if an individual is mosaic (germline or somatic), the original population of cells from which the DNA is derived may contain a mixture of diploid or haploid cells of the same or different genotypes. In some embodiments, the mixture of DNA molecules may be derived from DNA extracted from a single cell. In some embodiments, the mixture of DNA molecules may be derived from DNA extracted from a mixture of two or more cells of the same individual or two or more cells of different individuals. In some embodiments, the mixture of DNA molecules may be derived from DNA isolated from a biological material that is already free of cells, such as blood plasma, which is known to contain cell-free DNA. In some embodiments, the biological material may be a mixture of DNA from one or more individuals, as in the case of pregnancy, where fetal DNA has been shown to be present in the mixture. In some embodiments, the biological material may be derived from a mixture of cells found in maternal blood, some of which are of fetal origin. In some embodiments, the biological material may be cells from blood during pregnancy enriched in fetal cells.

Circularizing Probes Some embodiments of the present disclosure include the use of "ligated inverted probes" (LIPs), previously described in the literature, to amplify target loci either before or after amplification with non-LIP primers in the multiplex PCR methods of the present invention. LIPs is a generic term meant to encompass techniques that involve creating circular DNA molecules, where the probes are designed to hybridize to regions of the target DNA on either side of the target allele, such that with the addition of the appropriate polymerase and/or ligase, and appropriate conditions, buffers and other reagents, the complementary inverted regions of DNA spanning the target allele are completed, creating a circular loop of DNA that captures the information found in the target allele. LIPs are also referred to as pre-circularized probes, pre-circularizing probes or circularizing probes. LIP probes can be linear DNA molecules between 50 and 500 nucleotides in length, and in some embodiments between 70 and 100 nucleotides in length, and in some embodiments can be longer or shorter than described herein. Other embodiments of the present disclosure involve different implementations of the LIP technology, such as Padlock probes and molecular inverted probes (MIPs).

One way to target a specific location for sequencing is to synthesize a probe whose 3' and 5' ends anneal to the target DNA in an inverted fashion at locations adjacent to and on either side of the target region, so that adding DNA polymerase and DNA ligase results in extension from the 3' end, adding bases to the single-stranded probe that is complementary to the target molecule (gap filling), and then ligating the new 3' end to the 5' end of the original probe, resulting in a circular DNA molecule that can then be isolated from background DNA. The probe ends are designed to be adjacent to the target region of interest. One aspect of this approach is commonly referred to as MIPS and has been used in conjunction with array technology to determine the nature of the filled sequence. One drawback to using MIPs in the context of measuring allele ratios is that the hybridization, circularization and amplification steps do not occur at equal rates for different alleles at the same locus. The result is a measurement of allele ratios that do not represent the actual allele ratios present in the original mixture.

In one embodiment, the circularization probe is constructed such that the region of the probe designed to hybridize upstream of the target polymorphic locus and the region of the probe designed to hybridize downstream of the target polymorphic locus are covalently connected through a non-nucleic acid backbone. This backbone can be any biocompatible molecule or combination of biocompatible molecules. Some examples of possible biocompatible molecules are poly(ethylene glycol), polycarbonate, polyurethane, polyethylene, polypropylene, sulfone polymers, silicone, cellulose, fluoropolymers, acrylic compounds, styrene block copolymers, and other block copolymers.

In an embodiment of the present disclosure, this technique is modified to be easily amenable to sequencing as a means of examining the filling in the sequence. In order to preserve the original allele ratio of the original sample, at least one important consideration must be taken into account. The variable position between different alleles in the gap-fill region should not be too close to the probe binding site, as there may be bias initiation by the DNA polymerase resulting in differentiation of the variants. Another consideration is that there may be additional variation in the probe binding site that correlates with the variants in the gap-fill region that may result in unequal amplification from the different alleles. In an embodiment of the present disclosure, the 3' and 5' ends of the pre-circularization probe are designed to hybridize to bases that are one or a few positions away from the variant position (polymorphic site) of the target allele. The number of bases between the polymorphic site (SNP or other type) and the base to which the 3' and/or 5' ends of the pre-circularization probe are designed to hybridize can be 1 base, 2 bases, 3 bases, 4 bases, 5 bases, 6 bases, 7-10 bases, 11-15 bases, or 16-20 bases, 20-30 bases, or 30-60 bases. Forward and reverse primers can be designed to hybridize different numbers of bases away from the polymorphic site. Circularization probes can be generated in large numbers using current DNA synthesis techniques, allowing very large numbers of probes to be generated and potentially pooled, allowing many loci to be interrogated simultaneously. Working with over 300,000 probes has been reported. Two articles that discuss methods involving circularized probes that can be used to measure genomic data of a target individual include Porreca et al., Nature Methods, 2007, vol. 4(11), pp. 931-936; and Turner et al., Nature Methods, 2009, vol. 6(5), pp. 315-316. The methods described in these articles can be used in combination with other methods described herein. Certain steps of the methods from these two articles can be used in combination with other steps from other methods described herein.

In some embodiments of the methods disclosed herein, the genetic material of the target individual is optionally amplified, then hybridized with a pre-circularized probe, gap-filling is performed to fill in the bases between the two ends of the hybridized probe, the two ends are ligated to form a circularized probe, and the circularized probe is amplified, for example, using rolling circle amplification. Once the genetic information of the desired target allele is captured in a properly designed circularized oligonucleotide probe, for example, a LIP system, the genetic sequence of the circularized probe can be measured to provide the desired sequence data. In some embodiments, a properly designed oligonucleotide probe can be directly circularized in the unamplified genetic material of the target individual and then amplified. It should be noted that any number of amplification procedures, including rolling circle amplification, MDA, or other amplification protocols, can be used to amplify the original genetic material or to circularize the LIP. Different methods can be used to measure genetic information on a target genome, for example, high-throughput sequencing, Sanger sequencing, other sequencing methods, capture by hybridization, capture by circularization, multiplex PCR, other hybridization methods, and combinations thereof.

Once an individual's genetic material has been measured using one or a combination of the above methods, informatics-based methods, such as the PARENTAL SUPPORT™ method, along with appropriate genetic measurements, it can then be used to determine the ploidy state of one or more chromosomes in the individual, and/or the genetic state of one or a set of alleles (particularly alleles that correlate with a disease or genetic condition of interest). It is noted that the use of LIP for multiplexed capture of genetic sequences and subsequent genotyping using sequencing has been reported. However, the use of sequencing data generated by LIP-based strategies to amplify genetic material found in single cells, small numbers of cells, or extracellular DNA has not been used to determine the ploidy state of a target individual.

The application of informatics-based methods to determine the ploidy status of individuals from genetic data measured by hybridization arrays, e.g., ILLUMINA INFINIUM arrays or AFFYMETRIX gene chips, has been described in references elsewhere in this document. However, the methods described herein represent an improvement over methods previously described in the literature. For example, LIP-based approaches followed by high-throughput sequencing unexpectedly provide better genotype data, as the approaches have greater capacity for multiplexing, greater capture specificity, greater uniformity, and less allelic bias. Greater multiplexing allows more alleles to be targeted, resulting in more accurate results. Greater uniformity allows more targeted alleles to be measured, resulting in more accurate results. Lower rates of allelic bias reduce the rate of false calls, resulting in more accurate results. More accurate results result in improved clinical outcomes and better medical care.

It is important to note that LIP can be used as a method to target specific loci in a sample of DNA for genotyping by methods other than sequencing. For example, LIP can be used to target DNA for genotyping using SNP arrays or other DNA or RNA based microarrays.

Ligation-Mediated PCR
Ligation-mediated PCR can be used to amplify target loci before or after PCR amplification with unligated primers. Ligation-mediated PCR is a method of PCR used to preferentially enrich a sample of DNA by amplifying one or more loci in a mixture of DNA, the method comprising the steps of obtaining a set of primer pairs, each primer of the pair containing a target specific sequence and a non-target sequence, preferably the target specific sequence is designed to anneal to a target region, one upstream and one downstream of the polymorphic site, and the target specific sequence is within 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11-20, 20-30, 30-40, 40-50, 50-60, 60-70, 70-80, 80-90, 90-100, 100-1100, 1100-1200, 1200-1300, 1300-1400, 1400-1500, 1500-2600, 1600-2700, 1700-3000, 1800-2900, 1900-4000, 2000-5000, 2100-2400, 2200-2500, 2300-2400, 2400-2500, 2500-3000, 2600-2700, 2700-3000, 2800-3000, 2900-4000, 3000-4000, 3100-4000, 3200-4000, 3300-4000, 3400-4000, 3500-4000 The steps may be separated by 1-30, 31-40, 41-50, 51-100, or more than 100, polymerizing DNA from the 3' end of the upstream primer to fill in the single-stranded region between it and the 5' end of the downstream primer with nucleotides complementary to the target molecule, ligating the last polymerized base of the upstream primer to the adjacent 5' base of the downstream primer, and amplifying only the polymerized and ligated molecules using a non-target sequence containing the 5' end of the upstream primer and the 3' end of the downstream primer. Primer pairs for distinct targets can be mixed in the same reaction. The non-target sequence serves as a universal sequence, so all successfully polymerized and ligated primer pairs can be amplified using a single pair of amplification primers.

Capture by Hybridization In some embodiments, the method of the present disclosure can include using any of the following hybridization capture methods in addition to amplifying the target loci using multiplex PCR: Preferential enrichment of a specific set of sequences in the target genome can be achieved in a number of ways. Elsewhere in this document, we describe how LIP can be used to target a specific set of sequences, but in all of these applications, other targeting and/or preferential enrichment methods can be used equally well for the same purpose. One example of another targeting method is capture by hybridization techniques. Some examples of commercial capture by hybridization techniques include AGILENT's SURE SELECT and ILLUMINA's TruSeq. In hybridization capture, a set of oligonucleotides that are complementary or nearly complementary to the sequences of the desired target are hybridized to a mixture of DNA and then allowed to be physically separated from the mixture. Once the desired sequence is hybridized with the targeting oligonucleotide, the action of physically removing the targeting oligonucleotide will also remove the target sequence. Once the hybridized oligos are removed, they can be heated above their melting temperature and amplified. Some methods for physically removing the targeting oligonucleotide are by covalently binding the targeting oligonucleotide to a solid support, such as a magnetic bead or chip. Another method for physically removing the targeting oligonucleotide is by covalently binding the targeting oligonucleotide to a molecular moiety that has a strong affinity for another molecular moiety. An example of such a molecular pair is biotin and streptavidin, for example, as used in SURE SELECT. Thus, the target sequence is covalently attached to a biotin molecule, and after hybridization, the biotinylated oligonucleotide with which the target sequence is hybridized can be pulled down using a solid support with streptavidin added.

Hybrid capture involves hybridizing a probe complementary to the target of interest to a target molecule. Hybrid capture probes were originally developed to target and enrich large portions of the genome with relative uniformity between targets. In that application, it was important that all targets amplified had sufficient uniformity that all regions could be detected by sequencing, but no attention was paid to preserving the proportion of alleles in the original sample. After capture, the alleles present in the sample can be determined by direct sequencing of the captured molecules. These sequencing reads can be analyzed and counted according to allele type. However, with current technology, the allele distribution of the measured captured sequences generally does not represent the original allele distribution.

In one embodiment, allele detection is performed by sequencing. To capture the identity of the allele at the polymorphic site, it is essential that the sequencing read spans the allele in question to assess the allelic composition of the captured molecule. Because the captured molecules are often of variable length, when sequenced, the entire molecule must be sequenced to ensure that the mutation location overlaps. However, cost considerations and technical limitations regarding the maximum possible length and accuracy of the sequencing reads make sequencing of the entire molecule impractical. In one embodiment, the length of the read can be increased from about 30 bases to about 50 or 70 bases, allowing a significant increase in the number of reads that overlap the mutation location in the target sequence.

Another way to increase the number of reads interrogating a locus of interest is to decrease the length of the probe, provided that it does not result in bias of the underlying enriched allele. The length of the synthesized probe should be long enough for two probes designed to hybridize with two different alleles found at one locus to hybridize with approximately equal affinity to the various alleles in the original sample. Currently, methods known in the art generally describe probes longer than 120 bases. In current embodiments, when the allele is one or a few bases, the capture probe may be less than about 110 bases, less than about 100 bases, less than about 90 bases, less than about 80 bases, less than about 70 bases, less than about 60 bases, less than about 50 bases, less than about 40 bases, less than about 30 bases, and less than about 25 bases, which is sufficient to ensure equal enrichment from all alleles. When the mixture of DNA to be enriched using hybrid capture techniques is a mixture containing free floating DNA isolated from blood, e.g., maternal blood, the average length of the DNA is fairly short, generally less than 200 bases. Using shorter probes increases the likelihood that the hybrid capture probe will capture the desired DNA fragment. Larger variations may require longer probes. In an embodiment, the variation of interest is one (SNP) to a few bases in length. In an embodiment, targeted regions within the genome can be preferentially enriched using hybrid capture probes, where the hybrid capture probes are less than 90 bases in length, and may be less than 80 bases, less than 70 bases, less than 60 bases, less than 50 bases, less than 40 bases, less than 30 bases, or less than 25 bases in length. In certain embodiments, to increase the likelihood that the desired allele will be sequenced, the length of the probe designed to hybridize to the region adjacent to the location of the polymorphic allele can be reduced from more than 90 bases to about 80 bases, or to about 70 bases, or to about 60 bases, or to about 50 bases, or to about 40 bases, or to about 30 bases, or to about 25 bases.

To allow capture, there is a minimum overlap between the synthesized probe and the target molecule. This synthesized probe can be as short as possible, but still be larger than this minimum required overlap. The effect of using a shorter probe length to target a polymorphic region is that more molecules will overlap with the target allele region. The fragmentation state of the original DNA molecule also influences the number of reads that overlap with the target allele. Some DNA samples, such as plasma samples, are already fragmented due to biological processes occurring in vivo. However, samples with longer fragments benefit from fragmentation prior to sequencing library preparation and enrichment. When both the probe and the fragments are short (approximately 60-80 bp), maximum specificity can be achieved with relatively few sequence reads that cannot overlap with the critical region of interest.

In some embodiments, the hybridization conditions can be adjusted to maximize uniformity in the capture of different alleles present in the original sample. In some embodiments, the hybridization temperature is reduced to minimize differences in hybridization bias between alleles. Methods known in the art avoid using lower temperatures for hybridization because reducing the temperature has the effect of increasing hybridization of the probe with unintended targets. However, if the goal is to preserve the allele ratio with maximum fidelity, a technique using a lower hybridization temperature will result in the most accurate allele ratio, despite the fact that current art teachings avoid this technique. Hybridization temperatures can also be increased to require greater overlap between the target and the synthesized probe, so that only targets with substantial overlap of the target region are captured. In some embodiments of the present disclosure, the hybridization temperature is reduced from the normal hybridization temperature to about 40°C, to about 45°C, to about 50°C, to about 55°C, to about 60°C, to about 65°C, or to about 70°C.

In some embodiments, a hybrid capture probe can be designed such that the region of the capture probe that has DNA complementary to the DNA found in the region adjacent to the polymorphic allele is not immediately adjacent to the polymorphic site. Instead, the capture probe can be designed such that the region of the capture probe that is designed to hybridize to the DNA adjacent to the target polymorphic site is separated from the portion of the capture probe that makes van der Waals contact with the polymorphic site by a small distance, the length of which is equal to one or a few bases. In some embodiments, a hybrid capture probe is designed to hybridize to a region adjacent to but not across the polymorphic allele, which is referred to as an adjacent capture probe. The length of the adjacent capture probe can be less than about 120 bases, less than about 110 bases, less than about 100 bases, less than about 90 bases, and can be less than about 80 bases, less than about 70 bases, less than about 60 bases, less than about 50 bases, less than about 40 bases, less than about 30 bases, or less than about 25 bases. The region of the genome targeted by the flanking capture probe may be separated from the polymorphic locus by 1 base pair, 2 base pairs, 3 base pairs, 4 base pairs, 5 base pairs, 6 base pairs, 7 base pairs, 8 base pairs, 9 base pairs, 10 base pairs, 11-20 base pairs, or more than 20 base pairs.

Description of targeted capture based disease screening tests using targeted sequence capture. Custom targeted sequence capture such as those currently offered by AGILENT (SURE SELECT), ROCHE-NIMBLEGEN or ILLUMINA. Capture probes can be custom designed to ensure capture of various types of mutations. For point mutations, one or more probes overlapping the point mutation should be sufficient to capture and sequence the mutation.

For small insertions or deletions, one or more probes overlapping the mutation may be sufficient to capture and sequence the fragment containing the mutation. Hybridization may be less efficient between the probes generally designed to reference the genomic sequence and limited capture efficiency. To ensure capture of the fragment containing the mutation, two probes can be designed, one matching the normal allele and one matching the mutant allele. Longer probes can enhance hybridization. Multiple overlapping probes can enhance capture. Finally, placing probes immediately adjacent but not overlapping the mutation may allow relatively similar capture efficiency of the normal and mutant alleles.

For simple tandem repeats (STRs), probes that overlap these highly variable sites are unlikely to capture the fragments successfully. To enhance capture, probes can be placed close to, but not overlapping, the variable sites. The fragments can then be sequenced as usual to reveal the length and composition of the STRs.

For large deletions, a series of overlapping probes, a common approach currently used in exon capture systems, may work. However, using this approach, it may be difficult to determine whether an individual is heterozygous. Targeting and evaluating SNPs within the captured region can potentially indicate loss of heterozygosity across the region, indicating that the individual is a carrier. In an embodiment, non-overlapping or singleton probes can be placed across the potentially deleted region, and the number of captured fragments can be used as a measure of heterozygosity. If an individual has a large deletion, it is expected that half the number of fragments will be available for capture compared to the non-deleted (diploid) reference locus. Thus, the number of reads obtained from the deleted region should be approximately half the number of reads obtained from the normal diploid locus. The sequencing read depth from multiple singleton probes across the potentially deleted region can be aggregated and averaged to increase the signal and improve diagnostic confidence. The two approaches can also be combined: targeting SNPs to identify loss of heterozygosity and using multiple singleton probes to obtain a quantitative measure of the amount of underlying fragment from the locus. Either or both of these strategies can be combined with the other to better achieve the same results.

During testing, detection of cfDNA in a male fetus, indicated by the presence of a Y chromosome fragment that is captured and sequenced in the same test, and either an X-linked dominant mutation in which the mother and father are unaffected or a dominant mutation in which the mother is unaffected, indicates an increased risk to the fetus. Detection of two mutant recessive alleles in the same gene in an unaffected mother means that the fetus has inherited a mutant allele from the father and potentially a second mutant allele from the mother. In all cases, follow-up testing with amniocentesis or chorionic villus sampling may also be indicated.

Targeted capture-based disease screening tests can be combined with targeted capture-based non-invasive prenatal diagnostic tests for aneuploidy. There are a number of ways to reduce the variability of the depth of read (DOR): for example, primer concentrations can be increased, longer targeted amplification probes can be used, or more STA cycles can be performed (e.g., more than 25, more than 30, more than 35, or even more than 40).

Exemplary Methods for Determining the Number of DNA Molecules in a Sample Described herein are methods for determining the number of DNA molecules in a sample by generating a uniquely identified molecule for each original DNA molecule in the sample during a first round of DNA amplification. Described herein are procedures that achieve the above objectives and are followed by single molecule or clonal sequencing.

The technique involves targeting one or more specific loci and generating tagged copies of the original molecule such that each targeted locus has a unique tag and can be distinguished from one another when the barcodes are sequenced using clonal or single molecule sequencing. Each unique sequenced barcode represents a unique molecule in the original sample. At the same time, the sequencing data is used to confirm the locus from which the molecule originates. This information can be used to determine the number of unique molecules in the original sample for each locus.

This method can be used for any application requiring a quantitative assessment of the number of molecules in the original sample. Additionally, the number of unique molecules of one or more targets can be related to the number of unique molecules relative to one or more other targets to determine relative copy numbers, allele distributions, or allele ratios. Alternatively, the number of copies detected from various targets can be modeled by distribution to identify the most likely number of copies of the original target. Applications include, but are not limited to, detection of insertions and deletions, such as those found in carriers of Duchenne muscular dystrophy; quantification of deletions or duplications of chromosomal segments, such as those observed in copy number variants; chromosome copy numbers in samples from born individuals; chromosome copy numbers in samples from unborn individuals, such as embryos or fetuses.

The method can be combined with the evaluation of coincident variations contained in the targeting by sequence. This can be used to determine the number of molecules representing each allele in the original sample. This copy number method can be combined with the evaluation of SNPs or other sequence variations to determine chromosomal copy numbers in born and unborn individuals; identification and quantification of copies from loci that have short sequence variations but in which they can be amplified from multiple target regions by PCR, for example for carrier detection of spinal muscular atrophy; determination of copy numbers of various sources of molecules from samples consisting of a mixture of different individuals, for example for detection of fetal aneuploidy from free floating DNA obtained from maternal plasma.

In an embodiment, a method relating to a single target locus may include one or more of the following steps: (1) designing a standard pair of oligomers for PCR amplification of a specific locus; (2) adding a sequence of specific bases with no or minimal complementarity to the target locus or genome to the 5' end of one of the target-specific oligomers during synthesis. This sequence, called the tail, is a known sequence that is used for subsequent amplification and is followed by a sequence of random nucleotides. These random nucleotides include random regions. The random regions include randomly generated sequences of nucleic acids that are stochastically different between each probe molecule. Thus, after synthesis, the tailed oligomer pool consists of a population of oligomers that start with a known sequence, followed by an unknown sequence that differs between molecules, followed by the target-specific sequence. (3) performing one round of amplification (denaturation, annealing, extension) using only the tailed oligomers. (4) Add an exonuclease to the reaction to effectively stop the PCR reaction and incubate the reaction at an appropriate temperature to remove the forward single-stranded oligo that did not anneal to the template and extend it to form a double-stranded product. (5) Incubate the reaction at an elevated temperature to denature the exonuclease and eliminate its activity. (6) Add a new oligonucleotide complementary to the tail of the oligomer used in the first reaction to the reaction along with other target-specific oligomers to allow PCR amplification of the product generated in the first round of PCR. (7) Continue amplification to generate enough product for downstream clonal sequencing. (8) Measure the amplified PCR products that have a sufficient number of bases in the sequence by a number of methods, for example, clonal sequencing.

In an embodiment, the disclosed method involves targeting multiple loci in parallel or otherwise. Primers for different target loci can be generated independently and mixed to create a multiplex PCR pool. In an embodiment, the original sample can be split into subpools, with different loci targeted in each subpool, followed by recombination and sequencing. In an embodiment, the pool can be tagged before being subdivided and several amplification cycles performed to ensure efficient targeting of all targets, then split, refined, and then amplified by continuing amplification using a smaller set of primers in the subdivided pool.

One example of an application in which this technology would be particularly useful is non-invasive prenatal aneuploidy diagnosis, where the ratio of alleles at a given locus or the distribution of alleles at several loci can be used to help determine the number of copies of a chromosome present in a fetus. In this situation, it is desirable to amplify the DNA present in the initial sample while maintaining the relative amounts of the various alleles. In some cases, especially when very little DNA is present, e.g., less than 5,000 copies of the genome, less than 1,000 copies of the genome, less than 500 copies of the genome, and less than 100 copies of the genome, a phenomenon called bottlenecking can occur. This means that there are a small number of copies of any given allele in the initial sample, and as a result of amplification bias, the amplified pool of DNA has a significantly different ratio of those alleles than the ratio of those alleles in the initial mixture of DNA. By applying a unique or nearly unique set of barcodes to each strand of DNA prior to standard PCR amplification, it is possible to eliminate n-1 copies of DNA from a set of n identical molecules of sequenced DNA originating from the same original molecule.

For example, consider a SNP in an individual's genome that is heterozygous, and a mixture of DNA from that individual where there are 10 molecules of each allele in the original sample of DNA. After amplification, there may be 100,000 molecules of DNA corresponding to that locus. Due to stochastic processes, the ratio of DNA may be anywhere from 1:2 to 2:1, but because each of the original molecules was tagged with a unique tag, it is possible to determine that the DNA in the amplified pool originates from exactly 10 molecules of DNA from each allele. Thus, this method provides a more accurate measure of the relative amount of each allele than methods that do not use this technique. This method provides accurate data for methods where it is desirable to minimize the relative amount of allele bias.

The association of the sequenced fragments with the target locus can be achieved in a number of ways. In one embodiment, a molecular barcode corresponding to the target sequence, as well as a sufficient number of unique bases are linked to the target fragment to obtain a sequence of sufficient length to allow unambiguous identification of the target locus. In another embodiment, the molecular barcoding primer containing the randomly generated molecular barcode can also contain a locus-specific barcode (locus barcode) that identifies the target with which it is associated. This locus barcode is identical among all molecular barcoding primers for each individual target, and therefore among all of the resulting amplification products, but is distinct from all other targets. In one embodiment, the tagging method described herein can be combined with a one-sided nesting protocol.

In an embodiment, the design and generation of molecular barcoding primers can be implemented as follows: the molecular barcoding primers may consist of a sequence that is not complementary to the target sequence, followed by a random molecular barcode region, followed by a target-specific sequence. The sequence 5' of the molecular barcode can be used for partial sequence PCR amplification and may contain sequences useful in converting the amplified products into a library for sequencing. The random molecular barcode sequences can be generated in a number of ways. In a preferred method, the molecular tagged primers are synthesized to include all four bases for reaction during synthesis of the barcode region. All of the bases or various combinations of bases can be specified using the IUPAC DNA ambiguity code. In this way, the population of molecules synthesized contains a random mixture of sequences within the molecular barcode region. The length of the barcode region determines how many primers contain unique barcodes. The number of unique sequences is related to the length of the barcode region as N ^L , where N is the number of bases, typically 4, and L is the length of the barcode. A 5-base barcode can provide up to 1024 unique sequences, and an 8-base barcode can provide 65536 unique barcodes. In an embodiment, the DNA can be measured by a sequencing method, where the sequence data represents the sequence of a single molecule. This can include direct sequencing of a single molecule, or a method referred to herein as clonal sequencing, where a single molecule is amplified to form a clone detectable by a sequencing instrument, but still represents a single molecule.

Representative Methods and Reagents for Quantification of Amplification Products Quantification of specific nucleic acid sequences of interest is typically performed by quantitative real-time PCR techniques such as TAQMAN (LIFE TECHNOLOGIES), INVADER probes (THIRD WAVE TECHNOLOGIES), etc. Such techniques suffer from a number of shortcomings, including limited capacity for parallel simultaneous analysis of multiple sequences (multiplexing) and the ability to generate accurate quantitative data only within a narrow range of possible amplification cycles (e.g., within the linear range of the logarithm of PCR amplification product yield vs. cycle number). DNA sequencing technologies, particularly high-throughput next-generation sequencing technologies (often referred to as massively parallel sequencing technologies), such as those employed in MYSEQ (ILLUMINA), HISEQ (ILLUMINA), ION TORRENT (LIFE TECHNOLOGIES), GENOME ANALYZERILX (ILLUMINA), GSFLEX+ (ROCHE454), can be used to quantitatively measure the number of copies of a sequence of interest present in a sample, thereby providing quantitative information about the starting material, such as copy number or transcription level. High-throughput genetic sequencers use barcoding (i.e., tagging samples with characteristic nucleic acid sequences) to identify specific samples from a population of individuals, thereby allowing the simultaneous analysis of multiple samples in a single run of the DNA sequencer. The number of times a given region of the genome in a library preparation (or other nucleic acid preparation of interest) is sequenced (number of reads) will be proportional to the number of copies (or, in the case of cDNA-containing preparations, the expression level) of that sequence in the genome of interest. However, preparation and sequencing of genetic libraries (and similar genome-derived preparations) can introduce many biases that interfere with obtaining accurate quantitative reads of the nucleic acid sequences of interest. For example, different nucleic acid sequences may have different amplification efficiencies during the nucleic acid amplification step that occurs during genetic library preparation or sample preparation.

Problems associated with differing amplification efficiencies can be mitigated by using certain embodiments of the present invention. The present invention includes various methods and compositions related to the use of inclusion criteria during the amplification process that can be used to improve quantification accuracy. As described herein and, inter alia, in U.S. Pat. Nos. 8,008,018; 7,332,277; WO 2012/078792 A2; and WO 2011/146632 A1, the present invention is particularly useful in the area of fetal aneuploidy detection by analyzing free floating fetal DNA in maternal blood. These patents are incorporated herein by reference in their entirety. Embodiments of the present invention are also useful for detection of aneuploidy in in vitro generated embryos. Commercially important detectable aneuploidies include aneuploidies of human chromosomes 13, 18, 21, X, and Y.

Embodiments of the invention can be used with human or non-human nucleic acids and are applicable to both animal and plant derived nucleic acids. Embodiments of the invention can also be used to detect and/or quantitate alleles associated with other genetic disorders characterized by deletions or insertions. Deletion-containing alleles may be detected in individuals suspected of being carriers of the allele of interest.

One embodiment of the present invention includes standards present in known amounts (relative or absolute). For example, consider a genetic library made from a genetic source in which chromosome 8 (containing locus A) is diploid and chromosome 21 (containing locus B) is triploid. A genetic library can be generated from a sample containing sequences in amounts that are a function of the number of chromosomes present in the sample, e.g., 200 copies of locus A and 300 copies of locus B. However, if locus A is amplified much more efficiently than locus B, there may be 60,000 copies of A amplicon and 30,000 copies of B amplicon after PCR, thus obscuring the chromosome copy number of the initial true genomic sample when analyzed by high-throughput DNA sequencing (or other quantitative nucleic acid detection techniques). To alleviate this problem, a standard sequence for locus A is employed, where the standard sequence is amplified with substantially the same efficiency as locus A. Similarly, a standard sequence for locus B is formed, where the standard sequence is amplified with substantially the same efficiency as locus B. Prior to PCR (or other amplification technique), a reference sequence for locus A and a reference sequence for locus B are added to the mixture. These reference sequences are present in known amounts, either as relative or absolute amounts. Thus, under the same set of conditions, if a 1:1 mixture of reference sequence A and reference sequence B were added to the mixture in the previous example (prior to amplification), 3000 copies of reference A amplicon would be produced and 1000 copies of reference B amplicon would be produced, indicating that locus A is amplified 3 times more efficiently than locus B.

One or more selected regions of the genome containing the SNPs (or other polymorphisms) of interest can be specifically amplified and subsequently sequenced. This target specific amplification can be performed during the formation of a genetic library for sequencing. The library can include many target amplified regions. In some embodiments, the library includes at least 10; 100; 500; 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 regions of interest. Examples of such libraries are described herein and can also be found in U.S. Patent Publication No. 2012/0270212, filed November 18, 2011, which is incorporated herein by reference in its entirety.

Many high-throughput DNA sequencing techniques require modification of the starting genetic material for library formation, e.g., ligation of universal priming sites and/or barcodes, to facilitate clonal amplification of small nucleic acid fragments prior to subsequent sequencing reactions. In some embodiments, one or more reference sequences are added during genetic library formation or to precursor components of the genetic library prior to library amplification. The reference sequence can be selected to mimic (yet be distinguishable based on nucleotide base sequence) the target genomic fragments prepared for sequencing by the high-throughput genetic sequencing technique. In one embodiment, the reference sequence can be identical to the target genomic fragment except for 1, 2, 3, 4-10, or 11-20 nucleotides. In some embodiments, if the target genetic sequence contains a SNP, the reference sequence can be selected to be identical to the SNP except for the nucleotide of the polymorphic base, which is one of four nucleotides not observed at the natural site. The reference sequence can be used for highly multiplexed analysis of multiple target loci (e.g., polymorphic loci). The standard sequences can be added in known amounts (relative or absolute) during the process of library formation (pre-amplification) to provide a reference metric for greater accuracy in determining the amount of target sequences of interest in the analyzed sample. The combination of information on the known amount of standard sequences used, together with information on the previously characterized ploidy level, e.g., ploidy level library formation for sequencing formed from a genome where all autosomes are known to be diploid, can be used to calibrate the amplification characteristics of each standard sequence for its corresponding target sequence and can take into account the variation between batch mixtures containing multiple standard sequences. Given that it is often necessary to analyze a large number of loci simultaneously, it is useful to generate mixtures containing many sets of standard sequences. An embodiment of the present invention includes mixtures containing multiple standard sequences. Ideally, the amount of each standard sequence in the mixture would be known with high accuracy. However, this ideal is extremely difficult to achieve. This is because, in practice, there is a large variation in the amount of each standard sequence in a mixture, especially in the amount of standard sequences in a mixture containing a large number of different synthetic oligonucleotides. This variation can come from many sources, including batch-to-batch variation in in vitro oligonucleotide synthesis efficiency, volumetric inaccuracies, and pipetting variations. Moreover, this variation can occur even between batches that theoretically contain exactly the same set of standard sequences in exactly the same amounts. Therefore, it is meaningful to independently calibrate each batch of standard sequences. The batches of standard sequences can be calibrated against a reference genome of known chromosomal composition. The batches of standard sequences can be calibrated by sequencing the batches of standard sequences with minimal or no amplification steps included in the sequencing protocol. Embodiments of the invention include calibrated mixtures of different standard sequences. Other embodiments of the invention include methods of calibrating mixtures of different standard sequences and calibrated mixtures of different standard sequences produced by the methods.

Various embodiments of the target reference sequence mixture and their methods of use may include at least 10; 100; 500; 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 or more reference sequences, as well as various intermediate numbers. The number of reference sequences may be the same as the number of target sequences selected for analysis during the generation of the target library for DNA sequencing. However, in some embodiments, it may be advantageous to use a number of reference sequences that is less than the number of target regions in the library to be constructed. It may be advantageous to use a smaller number to avoid reaching the limits of the sequencing capacity of the high-throughput DNA sequencer employed. The number of standard sequences may be 50% or less of the number of target regions, 40% or less of the number of target regions, 30% or less of the number of target regions, 20% or less of the number of target regions, 10% or less of the number of target regions, 5% or less of the number of target regions, 1% or less of the number of target regions, as well as various intermediate numbers. For example, if a gene library is created using 15,000 pairs of primers targeting specific SNP-containing loci, an appropriate mixture containing 1500 standard sequences corresponding to 1500 of the 15,000 target loci can be added prior to the amplification step of library construction.

The amount of standard sequence added during library construction may vary widely between individual embodiments. In some embodiments, the amount of each standard sequence may be approximately the same as the predicted amount of the target sequence present in the genomic material sample used for library preparation. In other embodiments, the amount of each standard sequence may be greater or less than the predicted amount of the target sequence present in the genomic material sample used for library preparation. The initial relative amounts of target and standard sequences are not critical for purposes of the present invention, but are preferably in the range of 100-fold greater to 100-fold less than the amount of the target sequence present in the genomic material sample used for library preparation. Excessive amounts of standard sequences may use too much of the sequencing capacity of the DNA sequencer for a given number of runs of the instrument. Using too little standard sequence will result in insufficient data to be useful in analyzing variations in amplification efficiency.

The reference sequence can be selected to be a nucleotide base sequence closely similar to the region of interest to be amplified, and preferably the reference sequence has exactly the same primer binding sites as the genomic region to be analyzed, i.e., the "target sequence". The reference sequence must be distinguishable from the corresponding target sequence at a given locus. For convenience, this distinguishable region of the reference sequence is called the "marker sequence". In some embodiments, the marker sequence region of the target sequence contains a polymorphic region, e.g., a SNP, and may have primer binding regions located at both sides. The reference sequence can be selected to closely match the GC content of the corresponding target sequence. In some embodiments, the primer binding regions of the reference sequence are flanked on both sides by universal priming sites. These universal priming sites are selected to match the universal priming sites used in the genomic library to be analyzed. In other embodiments, the reference sequence is free of universal priming sites, and universal priming sites are added during library formation. The reference sequence is usually provided in single-stranded form. The reference sequence is defined relative to the corresponding target sequence, and the target sequence is amplified using sequence-specific reagents. In some embodiments, the target sequence includes a polymorphism of interest, e.g., a SNP, deletion, or insertion, present in the nucleic acid sample for analysis. The reference sequence is a synthetic polynucleotide that is similar in nucleotide base sequence to the target sequence, but is still distinguishable from the target sequence due to at least one nucleotide base difference, thereby providing a mechanism for distinguishing an amplified product sequence derived from the reference sequence from an amplified product sequence derived from the target sequence. The reference sequence is selected to have substantially the same amplification characteristics as the corresponding target sequence when amplified with the same set of amplification reagents, e.g., PCR primers. In some embodiments, the reference sequence can have the same primer sequence binding site as the corresponding target sequence. In other embodiments, the reference sequence can have a different primer sequence binding site than the corresponding target sequence. In some embodiments, the reference sequence can be selected to generate an amplified product of the same length as the amplified product from the corresponding target sequence. In other embodiments, the reference sequence can be selected to generate an amplified product of a slightly different length than the amplified product from the corresponding target sequence.

After the amplification reaction is completed, the library is sequenced on a high-throughput DNA sequencer, where individual molecules are clonally amplified and sequenced. The number of sequence reads for each allele of the target sequence is counted, as well as the number of sequence reads for the reference sequence corresponding to the target sequence. This process is also performed for at least another pair of target sequences and corresponding reference sequences. For example, for locus A, an X _A1 read is generated for allele 1 of locus A, an X _A2 read is generated for allele 2 of locus A, and an X _AC read is generated for reference sequence A. For each locus of interest, the ratio of (X _A1 plus X _A2 ) to X _AC is determined. As discussed previously, this process can be performed on a reference genome, for example, a genome in which all chromosomes are known to be diploid. This process can be repeated multiple times to obtain a large number of read values and measure the average number of reads and the standard deviation of the number of reads. This process is performed on a mixture containing many different reference sequences corresponding to different loci. The relative amounts of different standard sequences in a multiplex standard mixture can be determined by assuming (1) that X _A1 plus X _A2 corresponds to a known number of chromosomes, e.g., 2 for a normal human female genome, and (2) that the standard sequences have similar amplification (and detection) properties as their corresponding native loci. The calibrated multiplex standard sequence mixture can then be used to adjust for variations in amplification efficiency between different loci in a multiplex amplification reaction.

Other embodiments of the invention include methods and compositions for measuring copy number of a specific gene of interest, such as a mutant gene characterized by large deletions that can interfere with duplication and quantification by sequencing. Sequencing can have difficulty detecting alleles with such deletions. An amplification process that includes a reference sequence can be used to reduce this problem.

In one embodiment of the present invention, the target sequences for analysis are wild type (i.e., functional) and mutant forms of a gene characterized by a deletion. A representative such gene is SMN1, an allele with a deletion that is responsible for the genetic disease spinal muscular atrophy (SMA). It is worthwhile to detect individuals carrying mutant forms of the gene by high-throughput gene sequencing techniques. The application of such techniques to the detection of deletion mutations can be problematic, especially because of the lack of sequence observed during sequencing (as opposed to the detection of simple point mutations or SNPs). In such an embodiment, (1) a pair of amplification primers specific to the gene of interest that amplifies the gene of interest (or a portion thereof) and does not significantly amplify the mutant allele, (2) a reference sequence that corresponds to the wild type allele of the gene of interest (i.e., the target sequence) but differs by at least one detectable nucleotide base, (3) a pair of amplification primers specific to a second target sequence that serves as a reference sequence, and (4) a reference sequence that corresponds to the reference sequence.

In one embodiment of the present invention, a method is provided for determining the copy number of a gene of interest, the gene of interest having one intended allele that contains a deletion. The method can employ amplification reagents, e.g., PCR primers, that are specific to the gene of interest in that they do not amplify the deletion-containing allele of the gene of interest, but rather amplify at least a portion of the gene of interest, or the entire gene of interest, or a region adjacent to the gene of interest. Additionally, the method employs a reference sequence corresponding to the gene of interest, where the reference sequence differs from the gene of interest by at least one nucleotide base (so that the sequence of the reference sequence is easily distinguishable from the native gene of interest). Typically, the reference sequence contains the same primer binding sites as the gene of interest, thereby minimizing any amplification differences between the gene of interest and the reference sequence corresponding to the gene of interest. The reaction also includes amplification reagents specific to the reference sequence. The reference sequence is a sequence of known (or at least assumed to be known) copy number in the genome being analyzed. The reaction further includes a reference sequence corresponding to the reference sequence. Typically, a standard sequence that corresponds to a reference sequence contains the same primer binding sites as the reference sequence, thereby minimizing any amplification differences between the reference sequence and the standard sequence that corresponds to the reference sequence.

Representative Nucleic Acid Samples In some embodiments, a genetic sample can be prepared and/or purified. There are a number of standard procedures known in the art to accomplish such a goal. In some embodiments, the sample can be centrifuged to separate the various layers. In some embodiments, filtration can be used to isolate the DNA. In some embodiments, preparation of the DNA can involve amplification, separation, chromatographic purification, liquid-liquid separation, isolation, preferential enrichment, preferential amplification, targeted amplification, or any of a number of other techniques known in the art or described herein.

In some embodiments, the methods disclosed herein can be used in situations where there is very little DNA present, such as in in vitro fertilization or forensic situations where one or a few cells (typically less than 10 cells, less than 20 cells, or less than 40 cells) are available. In these embodiments, the methods disclosed herein are useful for making ploidy calls from small amounts of DNA where there is no other DNA contamination, but the small amount of DNA makes ploidy calls very difficult. In some embodiments, the methods disclosed herein can be used in situations where the target DNA is contaminated with the DNA of another individual, such as in maternal blood in the context of prenatal testing, paternity testing, or products of conception testing. Some other situations where these methods are particularly advantageous are in the case of cancer testing where only one or a few cells are present among a much larger amount of normal cells. The genetic measurements used as part of these methods can be performed on any sample containing DNA or RNA, including but not limited to blood, plasma, bodily fluids, urine, hair, tears, saliva, tissue, skin, fingernails, blastomeres, embryos, amniotic fluid, chorionic villus samples, feces, bile, lymph, cervical mucus, semen, or other cells or materials containing nucleic acids. In an embodiment, the methods disclosed herein can be performed together with a nucleic acid detection method, such as sequencing, microarray, qPCR, digital PCR, or other methods used to measure nucleic acids. If found desirable for any reason, the ratio of allele count probabilities at loci can be calculated, and the allele ratios can be used in combination with some of the methods described herein to determine ploidy state, as long as the methods are compatible. In some embodiments, the methods disclosed herein include a step of calculating the allele ratios at multiple polymorphic loci by computer from DNA measurements performed on the processed sample. In some embodiments, the methods disclosed herein, together with any combination of other improvements described in this disclosure, include a step of calculating by a computer allele ratio at a plurality of polymorphic loci from DNA measurements made on the processed sample.

In some embodiments, the method can be used to genotype a single cell, a small number of cells, 2-5 cells, 6-10 cells, 10-20 cells, 20-50 cells, 50-100 cells, 100-1,000 cells, or small amounts, e.g., 1-10 picograms, 10-100 picograms, 100 picograms to 1 nanogram, 1-10 nanograms, 10-100 nanograms, or 100 nanograms to 1 microgram of extracellular DNA.

Exemplary RNA Expression Surveys The multiplex PCR methods of the present invention can be used to increase the number of target loci that can be evaluated during gene expression profiling experiments. For example, the expression levels of thousands of genes can be monitored simultaneously to determine whether they have sequences (e.g., polymorphisms or other mutations) associated with disease (e.g., cancer) or increased risk of disease. These methods can be used to identify sequences (e.g., polymorphisms or other mutations) that are associated with increased or decreased risk of disease, e.g., cancer, by comparing gene expression (e.g., expression of specific mRNA alleles) in samples from patients with or without disease. In addition, the effect of a particular treatment, disease, or developmental stage on gene expression can be determined. Similarly, these methods can be used to identify which genes have changed expression in response to a pathogen or other organism by comparing gene expression in infected and uninfected cells or tissues. In these methods, the number of sequencing reads can be adjusted based on the frequency of the polymorphism to be detected (if the polymorphism being analyzed is present) so that sufficient reads are performed for the polymorphism to be detected.

In some embodiments, a sample containing RNA (e.g., mRNA) is amplified using reverse transcriptase (RT) and the resulting DNA (e.g., cDNA) is then amplified using DNA polymerase (PCR). The RT and PCR steps can be performed sequentially in the same reaction volume or separately. Any of the primer libraries of the present invention can be used in this reverse transcription polymerase chain reaction (RT-PCR) method. In various embodiments, reverse transcription can be performed using oligo-dT, random primers, a mixture of oligo-dT and random primers, or target locus-specific primers. To avoid amplification of contaminating genomic DNA, primers for RT-PCR can be designed such that a portion of one primer hybridizes to the 3' end of one exon and a portion of the other primer hybridizes to the 5' end of the adjacent exon. Such primers anneal to cDNA synthesized from spliced mRNA but not to genomic DNA. To detect amplification of contaminating DNA, RT-PCR primer pairs can be designed to flank a region that contains at least one intron. Products amplified from cDNA (without introns) are smaller than those amplified from genomic DNA (with introns). The size difference in the products is used to detect the presence of contaminating DNA. In some embodiments, when only the mRNA sequence is known, primer annealing sites are selected that are at least 300-400 base pairs apart, since fragments of this size from eukaryotic DNA are likely to contain splice junctions. Alternatively, the sample can be treated with DNase to degrade contaminating DNA.

Exemplary Paternity Testing Methods The accuracy of paternity testing can be improved using the multiplex PCR methods of the present invention because a large number of target loci can be analyzed at once (see, e.g., U.S. Patent Publication No. 2012/0122701, filed December 22, 2011, which is incorporated herein by reference in its entirety). For example, multiplex PCR methods allow thousands of polymorphic loci (e.g., SNPs) to be analyzed for use in the PARENTAL SUPPORT algorithm described herein to determine whether an alleged father is the biological father of a fetus. In some embodiments, the method includes the steps of: (i) simultaneously amplifying a plurality of polymorphic loci, including at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different polymorphic loci, on genetic material from the alleged father to generate a first set of amplified products; (ii) simultaneously amplifying a corresponding plurality of polymorphic loci of a mixed DNA sample, including fetal DNA and maternal DNA, from a blood sample of the pregnant mother to generate a second set of amplified products; (iii) calculating on a computer a probability that the alleged father is the biological father of the fetus using genotypic measurements based on the first and second sets of amplified products; and (iv) establishing whether the alleged father is the biological father of the fetus using the calculated probability that the alleged father is the biological father of the fetus. In various embodiments, the method further comprises simultaneously amplifying corresponding polymorphic loci on the genetic material from the mother to generate a third set of amplified products, where genotype measurements based on the first, second, and third sets of amplified products are used to calculate the probability that the alleged father is the biological father of the fetus.

Exemplary Embryo Characterization and Selection Methods The multiplex PCR methods of the present invention can be used to improve the selection of embryos for in vitro fertilization by allowing thousands of target loci to be analyzed at once (see, e.g., U.S. Patent Publication No. 2011/0092763, filed May 27, 2008, and filed December 22, 2011, which is incorporated by reference in its entirety). For example, the multiplex PCR methods can allow the analysis of thousands of polymorphic loci (e.g., SNPs) for use in the PARENTAL SUPPORT algorithm described herein to select embryos for in vitro fertilization from a collection of embryos.

In some embodiments, the invention provides a method for estimating the relative likelihood of each embryo from a set of embryos to develop as desired. In some embodiments, the method includes contacting a sample from each embryo with a library of primers that simultaneously hybridize to at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci to generate a mixture of each embryo, each sample being obtained from one or more cells from the embryo. In some embodiments, each reaction mixture is subjected to primer extension reaction conditions to generate an amplification product. In some embodiments, the method includes determining in a computer one or more characteristics of at least one cell from each embryo based on the amplification product, and estimating in a computer the relative likelihood of each embryo to develop as desired based on the one or more characteristics of the at least one cell for each embryo. In some embodiments, the method includes determining at least one trait using an informatics-based method, such as the PARENTAL SUPPORT algorithm described herein. In some embodiments, the trait includes a ploidy state. In some embodiments, the trait is selected from the group consisting of aneuploid, euploid, mosaic, nullisomy, monosomy, uniparental disomy, trisomy, tetrasomy, type of aneuploidy, unmatched copy error trisomy, matched copy error trisomy, aneuploidy of maternal origin, aneuploidy of paternal origin, presence or absence of disease-linked genes, chromosomal identity of all euploid chromosomes, abnormal gene state, deletion or duplication, likelihood of trait, and combinations thereof. The trait may be associated with a chromosome selected from the group consisting of chromosome 1, chromosome 2, chromosome 3, chromosome 4, chromosome 5, chromosome 6, chromosome 7, chromosome 8, chromosome 9, chromosome 10, chromosome 11, chromosome 12, chromosome 13, chromosome 14, chromosome 15, chromosome 16, chromosome 17, chromosome 18, chromosome 19, chromosome 20, chromosome 21, chromosome 22, the X chromosome, or the Y chromosome, and combinations thereof.

Exemplary Prenatal Diagnostic Methods The multiplex PCR methods of the present invention can be used to improve prenatal diagnostic methods, such as determining the ploidy status of fetal chromosomes. More accurate determinations can be made given the large number of target loci that are simultaneously amplified.

In an embodiment, the disclosure provides an ex vivo method for determining the ploidy state of chromosomes in a gestating fetus from genotype data measured from a mixed sample of DNA (i.e., DNA from the fetus's mother and DNA from the fetus) and, optionally, from samples of genetic material from the mother and, optionally, from genetic material from the father as well, by using a joint distribution model to generate a set of predicted allele distributions for different possible ploidy states in the fetus, taking into account the genotype data of the parents, comparing the predicted allele distributions with the actual allele distributions measured in the mixed sample, and selecting the ploidy state whose predicted allele distribution pattern most closely matches the observed allele distribution pattern. In an embodiment, the mixed sample is derived from maternal blood or maternal serum or plasma. In an embodiment, the mixed sample of DNA can be preferentially enriched at target loci (e.g., multiple polymorphic loci). In an embodiment, the preferential enrichment is performed such that allele bias is minimized. In an embodiment, the present disclosure relates to a composition of DNA that is preferentially enriched to reduce allele bias at multiple loci. In an embodiment, the allele distribution(s) are measured by sequencing DNA from a mixed sample. In an embodiment, a joint distribution model assumes that alleles are distributed in a binomial manner. In an embodiment, a set of predicted joint allele distributions is generated for genetically linked loci, taking into account existing recombination frequencies from various sources, e.g., using data from the International HapMap Consortium.

In an embodiment, the present disclosure provides a method for non-invasive prenatal diagnosis (NPD), specifically a method for determining the aneuploidy status of a fetus by observing allele measurements at multiple polymorphic loci in genotype data measured on a DNA mixture, where certain allele measurements indicate an aneuploid fetus, while other allele measurements indicate a euploid fetus. In an embodiment, the genotype data is measured by sequencing on a DNA mixture derived from maternal plasma. In an embodiment, the DNA sample can be preferentially enriched for DNA molecules corresponding to the multiple loci for which the allele distribution is calculated. In an embodiment, a sample of DNA containing only or almost exclusively maternal genetic material is measured, and in some cases, a sample of DNA containing only or almost exclusively paternal genetic material is also measured. In an embodiment, the genetic measurements of one or both parents are used together with the estimated fetal fraction to generate multiple predicted allele distributions corresponding to different possible underlying genetic states in the fetus, which predicted allele distributions can be referred to as hypotheses. In an embodiment, maternal genetic data is not determined by measuring genetic material that is exclusively or nearly exclusively maternal in nature, but is estimated from genetic measurements made on maternal plasma that contains a mixture of maternal and fetal DNA. In some embodiments, the hypotheses may include the ploidy of the fetus at one or more chromosomes, which segments of which fetal chromosomes were inherited from which parent, and combinations thereof. In some embodiments, the ploidy state of the fetus is determined by comparing the observed allele measurements with different hypotheses, at least some of which correspond to different ploidy states, and selecting the ploidy state that corresponds to the hypothesis that is most likely to be true given the observed allele measurements. In an embodiment, the method involves the use of allele measurement data from some or all of the measured SNPs, regardless of whether the locus is homozygous or heterozygous, and therefore does not involve the use of alleles at loci that are only heterozygous. This method may not be suitable for situations where the genetic data pertains to only one polymorphic locus. This method is particularly advantageous when the genetic data includes data for more than 10 polymorphic loci for the target chromosome, or more than 20 polymorphic loci. This method is particularly advantageous when the genetic data includes data for more than 50 polymorphic loci for the target chromosome, more than 100 polymorphic loci, or more than 200 polymorphic loci for the target chromosome. In some embodiments, the genetic data may include data for more than 500 polymorphic loci for the target chromosome, more than 1,000 polymorphic loci, more than 2,000 polymorphic loci, or more than 5,000 polymorphic loci for the target chromosome.

In an embodiment, the method disclosed herein provides a quantitative measure of the number of independent observations of each allele at a polymorphic locus. This differs from the majority of methods, such as microarrays or qualitative PCR, which provide information on the ratio of two alleles but do not quantify the number of independent observations of either allele. Methods that provide quantitative information on the number of independent observations only use the ratio to calculate ploidy, while the quantitative information is not useful in itself. To illustrate the importance of retaining information on the number of independent observations, consider a sample locus with two alleles, A and B. In the first experiment, 20 alleles A and 20 alleles B are observed, and in the second experiment, 200 alleles A and 200 alleles B are observed. In both experiments, the ratio (A/(A+B)) is equal to 0.5, but the second experiment conveys more information about the certainty of the frequency of alleles A or B than the first experiment. Some methods by other researchers involve averaging or summing the allelic ratios (channel ratios) (i.e., x _i /y _i ) from individual alleles and analyzing this ratio either by comparing it to a reference chromosome or by using rules on how the ratio is expected to behave in a particular situation. Such methods do not involve allele weighting and assume that one can ensure approximately the same amount of PCR product for each allele and that all alleles should behave in the same way. Such methods have a number of disadvantages and, more importantly, prevent the use of a number of improvements described elsewhere in this disclosure.

In an embodiment, the methods disclosed herein explicitly model the allele frequency distributions expected in disomy as well as the multiple allele frequency distributions that may be expected in the case of trisomy resulting from nondisjunction during meiosis I, nondisjunction during meiosis II, and/or nondisjunction during early mitosis of fetal development. To illustrate why this is important, consider the case where there is no crossover: nondisjunction during meiosis I results in trisomy with two different homologs inherited from one parent, in contrast to nondisjunction during meiosis II or during early mitosis of fetal development that results in two copies of the same homolog from one parent. Each scenario will result in different expected allele frequencies at each polymorphic locus and, due to genetic linkage, at all loci considered jointly. Crossover, which results in the exchange of genetic material between homologs, introduces a more complex inheritance pattern, which in an embodiment the method accommodates by using information on recombination rates in addition to the physical distance between loci. In one embodiment, to allow for improved discrimination between meiosis I nondisjunction and meiosis II or mitotic nondisjunction, the method incorporates into the model a probability of crossing over that increases with increasing distance from the centromere. Meiosis II and mitotic nondisjunction can be distinguished by the fact that mitotic nondisjunction generally results in identical or nearly identical copies of one homolog, whereas the two homologs present after a meiosis II nondisjunction event often differ due to one or more crossing overs during gamete formation.

In some embodiments, the methods disclosed herein include comparing the observed allele measurements to theoretical hypotheses corresponding to possible fetal genetic aneuploidies, and do not include quantifying the ratio of alleles at heterozygous loci. When the number of loci is less than about 20, ploidy determinations performed using a method including quantifying the ratio of alleles at heterozygous loci and ploidy determinations performed using a method including comparing the observed allele measurements to theoretical hypotheses of allele distribution corresponding to possible fetal genetic conditions may yield similar results. However, when the number of loci is more than 50, these two methods may yield significantly different results, and when the number of loci is more than 400, more than 1,000, or more than 2,000, these two methods are increasingly likely to yield significantly different results. These differences arise from the fact that methods that involve quantifying allele ratios at heterozygous loci without measuring the magnitude of each allele independently and aggregating or averaging the ratios preclude the use of techniques including using joint distribution models, performing linkage analysis, using binomial distribution models, and/or other advanced statistical techniques, whereas methods that involve comparing observed allele measurements to theoretical allele distribution hypotheses corresponding to possible fetal genetic states allow the use of these techniques that can substantially increase the accuracy of the determination.

In an embodiment, the method disclosed herein includes determining whether the distribution of observed allele measurements indicates a euploid or aneuploid fetus using a joint distribution model. The use of a joint distribution model is different from and a significant improvement over methods that determine heterozygosity rates by independently treating polymorphic loci in that the accuracy of the determinations obtained is significantly higher. Without being bound to any particular theory, it is believed that one reason for their high accuracy is that the joint distribution model takes into account the linkage between SNPs and the likelihood of crossovers that occurred during meiosis that gave rise to the gametes that form the embryo that will develop into the fetus. The purpose of using the concept of linkage in generating the predicted distribution of allele measurements for one or more hypotheses is to thereby enable the generation of predicted allele measurement distributions that correspond significantly better to reality than would be the case without the use of linkage. For example, consider two SNPs, 1 and 2, located near each other, and the mother has A at SNP1 and A at SNP2 on one homolog, and B at SNP1 and B at SNP2 on homolog 2. If the father has A for both SNPs on both homologs, and B is measured for SNP1 in the fetus, this indicates that homolog 2 has been inherited by the fetus, and therefore there is a much higher likelihood that a B is present at SNP2 in the fetus. Models that take linkage into account would predict this, but models that do not would not. Alternatively, if the mother has AB at SNP1 and AB at nearby SNP2, two hypotheses can be used that correspond to maternal trisomy at that location - one with a concordant copy error (nondisjunction in meiosis II or early fetal mitosis) and one with a discordant copy error (nondisjunction in meiosis I). In the case of a matching copy error trisomy, if the fetus inherited AA from the mother at SNP1, it is much more likely to inherit either AA or BB from the mother at SNP2, but not AB. In the case of a nonmatching copy error, the fetus will inherit AB from the mother at both SNPs. These predictions are made because of the allele distribution assumptions made by ploidy calling methods that take linkage into account, and therefore correspond to the actual allele measurements to a much greater extent than ploidy calling methods that do not take linkage into account. Note that linkage approaches are not possible when using methods that rely on calculating allele ratios and aggregating those allele ratios.

One reason that the accuracy of ploidy determination using a method that includes comparing observed allele measurements to theoretical hypotheses corresponding to possible fetal genetic states is believed to be higher is that when sequencing is used to measure alleles, the method can glean more information from data from alleles when the total number of reads is smaller than other methods, for example, methods that rely on calculating and summing allele ratios produce disproportionately weighted stochastic noise. For example, consider a case that involves measuring alleles using sequencing and where there is a set of loci where only five sequence reads were detected for each locus. In an embodiment, for each of the alleles, the data can be compared to the hypothesized allele distribution and weighted according to the number of sequence reads, and thus the data from these measurements are appropriately weighted and incorporated into the overall determination. This is in contrast to methods that involve quantifying allele ratios at heterozygous loci, which can only calculate 0%, 20%, 40%, 60%, 80% or 100% possible allele ratios, none of which come close to the expected allele ratio. In this latter case, the calculated allele ratio must be rejected due to insufficient reads or is disproportionately weighted, introducing stochastic noise into the decision, thereby reducing its accuracy. In an embodiment, each allele measurement can be treated as an independent measurement, in which case the relationship between measurements made on alleles at the same locus is no different from the relationship between measurements made on alleles at different loci.

In an embodiment, the method disclosed herein involves determining whether the distribution of observed allele measurements indicates a euploid or aneuploid fetus without including a step of comparing any metric to the allele measurements observed in a reference chromosome expected to be disomic (referred to as the RC method). This is a significant improvement over methods such as those using shotgun sequencing, which detect aneuploidy by assessing the proportion of randomly sequenced fragments from a suspect chromosome compared to one or more suspected disomic reference chromosomes. This RC method produces inaccurate results when the suspected disomic reference chromosome is not actually disomic. This can occur when the aneuploidy is more substantial than a single chromosome trisomy or when the fetus is triploid and all autosomes are trisomic. In the case of a female triploid (69, XXX) fetus, there are in fact no disomic chromosomes at all. The method described herein does not require a reference chromosome and can accurately identify trisomic chromosomes in a female triploid fetus. For each chromosome, hypothesis, child fraction, and noise level, a joint distribution model can be fitted without reference chromosome data, estimates of the overall child fraction, or a fixed reference hypothesis.

In an embodiment, the methods disclosed herein demonstrate how the observed allele distributions at polymorphic loci can be used to determine the ploidy state of the fetus with greater accuracy than prior art methods. In an embodiment, the method uses targeted sequencing to obtain mixed maternal-fetal genotypes at multiple SNPs, and optionally maternal and/or paternal genotypes, to first establish various expected allele frequency distributions under different hypotheses, then observe the quantitative allelic information obtained in the maternal-fetal mixtures, and evaluate which hypothesis best fits the data, and call the genetic state corresponding to the hypothesis that best fits the data as the correct genetic state. In an embodiment, the methods disclosed herein also use the degree of fit to generate a confidence that the called genetic state is the correct genetic state. In an embodiment, the method disclosed herein involves using an algorithm to analyze the distribution of alleles found for loci with different parental statuses, and comparing the observed allele distributions to expected allele distributions for different ploidy states for different parental statuses (different parental genotype patterns). This is different from and an improvement over methods that do not use methods that can estimate the number of independent instances of each allele at each locus in a mixed maternal-fetal sample. In an embodiment, the method disclosed herein involves determining whether the distribution of observed allele measurements indicates a euploid or aneuploid fetus using the observed allele distributions measured at loci where the mother is heterozygous. This is different from and an improvement over methods that do not use the observed allele distributions at loci where the mother is heterozygous, as it allows for approximately twice as much genetic measurement data from the set of sequence data to be used in determining ploidy, resulting in a more accurate determination, when DNA is not preferentially enriched or is preferentially enriched for loci that are not known to be highly informative for that particular target individual.

In an embodiment, the methods disclosed herein use a joint distribution model that assumes that the allele frequency at each locus is polynomial in nature (and therefore binomial if the SNP is biallelic). In some embodiments, the joint distribution model uses a beta-binomial distribution. If a measurement technique such as sequencing is used that provides a quantitative measure for each allele present at each locus, the binomial model can be applied to each locus and the underlying degree of allele frequency and the confidence in that frequency can be ascertained. The certainty of the observed ratios cannot be ascertained using methods known in the art that generate ploidy calls from allele ratios or methods in which quantitative allele information is discarded. The method is different from and an improvement over methods that calculate allele ratios and aggregate the ratios to make ploidy calls, since any method that involves calculating allele ratios at a particular locus and then aggregating those ratios necessarily assumes that the measured intensities or counts that indicate the amount of DNA from any given allele or locus are distributed in a Gaussian manner. The methods disclosed herein do not involve calculating allele ratios. In some embodiments, the methods disclosed herein may include incorporating the number of observations of each allele at multiple loci into the model. In some embodiments, the methods disclosed herein may include calculating the expected distribution itself, which allows for the use of a joint binomial distribution model, which may be more accurate than any model that assumes a Gaussian distribution of allele measurements. The likelihood that the binomial distribution model is significantly more accurate than the Gaussian distribution increases as the number of loci increases. For example, when examining less than 20 loci, the likelihood that the binomial distribution model is significantly superior is low. However, when using more than 100, or especially more than 400, or especially more than 1,000, or especially more than 2,000 loci, the likelihood that the binomial distribution model is significantly more accurate than the Gaussian distribution model becomes very high, thereby resulting in a more accurate ploidy determination. The likelihood that the binomial model is significantly more accurate than the Gaussian distribution model similarly increases as the number of observations at each locus increases. For example, when less than 10 distinct sequences are observed at each locus, the likelihood that the binomial model is significantly better is low. However, when more than 50 sequence reads are used for each locus, or especially more than 100 sequence reads, or especially more than 200 sequence reads, or especially more than 300 sequence reads, the likelihood that the binomial model is significantly more accurate than the Gaussian distribution model becomes very high, thereby resulting in more accurate ploidy determinations.

In an embodiment, the methods disclosed herein use sequencing to measure the number of instances of each allele at each locus in a DNA sample. Each sequencing read can be mapped to a specific locus and treated as a binary sequence read, or the probability of the identity of the read and/or mapping can be incorporated as part of the sequence read, resulting in a probabilistic sequence read, i.e., an estimated integer or fractional number of sequence reads that map to a given locus. Using binary counts or count probabilities, a binomial distribution can be used for each set of measurements, which allows for the calculation of confidence intervals around the number of counts. The ability to use a binomial distribution allows for more accurate ploidy estimates and more precise confidence intervals to be calculated. This is different from and an improvement over methods that use intensity to measure the amount of alleles present, such as using microarrays or using a fluorescent reader to measure the intensity of fluorescently tagged DNA in electrophoretic bands.

In one embodiment, the methods disclosed herein use aspects of the data set to determine parameters of the estimated allele frequency distribution for the data set. This is an improvement over methods that utilize a training or prior data set to set parameters for the expected allele frequency distribution or possibly the expected allele ratio. This is because there is a set of different conditions involved in the collection and measurement of any genetic sample, and therefore the method of using data from the data set to determine parameters of a joint distribution model to be used in determining ploidy for that sample is more likely to be accurate.

In an embodiment, the method disclosed herein involves using maximum likelihood to determine whether the distribution of observed allele measurements indicates a euploid or aneuploid fetus. Using maximum likelihood is different from and a significant improvement over single hypothesis rejection methods in that the accuracy of the resulting determination is significantly higher. One reason is that in single hypothesis rejection methods, the cutoff threshold is set based on only one measurement distribution rather than two, i.e., the threshold is usually not optimal. Another reason is that maximum likelihood methods allow the cutoff threshold to be optimized for each individual sample, rather than determining a cutoff threshold to be used for all samples regardless of the specific characteristics of each individual sample. Another reason is that using maximum likelihood methods allows a confidence to be calculated for each ploidy call. Being able to perform a confidence calculation for each call allows the practitioner to know which calls are accurate and which are more likely to be incorrect. In some embodiments, a wide variety of methods can be combined with maximum likelihood estimation to enhance the accuracy of ploidy calls. In an embodiment, the maximum likelihood method can be used in combination with the method described in U.S. Pat. No. 7,888,017. In an embodiment, the maximum likelihood method can be used in combination with a method using targeted PCR amplification to amplify the DNA in the mixed sample, followed by sequencing and analysis using a read counting method, such as TANDEM DIAGNOSTICS, presented at the International Congress of Human Genetics 2011, Montreal, October 2011. In an embodiment, the method disclosed herein involves estimating the fetal fraction of the DNA in the mixed sample, and using that estimate to calculate both the ploidy call and the confidence of the ploidy call. Note that this is different from and distinct from the method of using the estimated fetal fraction as a screen for sufficient fetal fraction, followed by a ploidy call using a single hypothesis rejection method that does not take the fetal fraction into account, nor does it result in a calculation of confidence for the call.

In an embodiment, the method disclosed herein takes into account the tendency of data to be noisy and contain errors by assigning a probability to each measurement. By using maximum likelihood to select the correct hypothesis from a set of hypotheses made using measurement data with assigned probabilistic estimates, it is more likely that inaccurate measurements will not be taken into account and accurate measurements will be used in the calculation leading to the ploidy call. To be more accurate, the method systematically reduces the influence of inaccurately measured data in the determination of ploidy. This is an improvement over methods where all data are assumed to be equally accurate or where outlying data are arbitrarily excluded from the calculation leading to the ploidy call. Current methods using channel ratio measurements claim to extend the method to a large number of SNPs by averaging individual SNP channel ratios. By not weighting individual SNPs by the expected variance of measurements based on the quality of the SNP and the observed read depth, the accuracy of the resulting statistics is reduced, which in turn significantly reduces the accuracy of the ploidy call, especially in borderline cases.

In certain embodiments, the methods disclosed herein do not presuppose knowledge of which SNPs or other polymorphic loci are heterozygous in the fetus. This method allows ploidy calls to be made when paternal genotype information is not available. This is an improvement over methods in which knowledge of which SNPs are heterozygous must be known in advance in order to appropriately select loci to target or to interpret genetic measurements obtained on a mixed fetal/maternal DNA sample.

The methods described herein are particularly advantageous when used with samples with small amounts of DNA available or with a low percentage of fetal DNA. This is due to the correspondingly high allele dropout rate that occurs when only small amounts of DNA are available and/or when the percentage of fetal DNA in the mixed fetal and maternal DNA sample is low. A high allele dropout rate, i.e., a large proportion of alleles were not measured for the target individual, leads to insufficiently accurate calculation of fetal fractions and insufficiently accurate ploidy determinations. The methods disclosed herein can use a joint distribution model that takes into account linkage in the inheritance pattern between SNPs, allowing for significantly more accurate ploidy determinations. The methods described herein allow for accurate ploidy determinations when the percentage of fetal DNA molecules in the mixture is less than 40%, less than 30%, less than 20%, less than 10%, less than 8%, or even less than 6%.

In some embodiments, it is possible to determine the ploidy state of an individual based on measurements of the individual's DNA when mixed with DNA of related individuals. In some embodiments, the mixture of DNA is free-floating DNA found in maternal plasma, which may include DNA from the mother with known karyotype and known genotype, and may also be mixed with fetal DNA with unknown karyotype and unknown genotype. With known genotype information from one or both parents, it is possible to predict multiple potential genetic states of the DNA in the mixed sample for different ploidy states, different chromosomal contributions from each parent to the fetus, and, optionally, different percentages of fetal DNA in the mixture. Each potential composition may be referred to as a hypothesis. The ploidy state of the fetus can then be determined by examining the actual measurements and determining which potential composition is most likely given the observed data.

Further discussion of the above points can be found elsewhere in this document.

Non-Invasive Prenatal Diagnosis (NPD)
The process of non-invasive prenatal diagnosis involves a number of steps. Some of the steps may include (1) obtaining genetic material from a fetus, (2) enriching fetal genetic material that may be present in a mixed sample ex vivo, (3) amplifying the genetic material ex vivo, (4) preferentially enriching specific loci of the genetic material ex vivo, (5) measuring the genetic material ex vivo, and (6) analyzing the genotype data ex vivo on a computer. Methods are described herein to reduce the implementation of these six and other related steps. At least some of the steps of the method are not directly applied to the body. In an embodiment, the present disclosure relates to methods of treatment and diagnosis applied to tissues and other biological materials isolated and separated from the body, at least some of the steps of the method are computer-implemented.

Some embodiments of the present disclosure allow clinicians to non-invasively determine the genetic status of a fetus during pregnancy, without endangering the health of the infant by harvesting genetic material from the fetus and without the mother having to undergo an invasive procedure. Furthermore, in certain aspects, the present disclosure allows the genetic status of the fetus to be determined with high accuracy, for example, significantly higher accuracy than the non-invasive maternal serum analyte-based screening of the triple test commonly used in prenatal care.

The high accuracy of the methods disclosed herein is a result of the informatics approaches described herein for analyzing genotype data. Modern technological advances have made it possible to measure large amounts of genetic information from genetic samples using methods such as high-throughput sequencing and genotyping arrays. The methods disclosed herein allow clinicians to take greater advantage of the large amounts of data available and to make more accurate diagnoses of fetal genetic conditions. Details of several embodiments are provided below. Different embodiments may include different combinations of the steps described above. Various combinations of different embodiments of different steps can be used interchangeably.

In an embodiment, a blood sample is obtained from a pregnant mother, and free floating DNA in the plasma of the mother's blood containing a mixture of both maternal and fetal DNA is isolated and used to determine the ploidy state of the fetus. In an embodiment, the method disclosed herein includes preferentially enriching DNA sequences in the mixture of DNA corresponding to polymorphic alleles such that the allele ratio and/or allele distribution remains largely unchanged upon enrichment. In an embodiment, the method disclosed herein involves highly efficient targeted PCR-based amplification such that a very high percentage of the resulting molecules correspond to the targeted locus. In an embodiment, the method disclosed herein includes sequencing the mixture of DNA containing both maternal and fetal DNA. In an embodiment, the method disclosed herein includes using the measured allele distribution to determine the ploidy state of the fetus while the mother is pregnant. In an embodiment, the method disclosed herein includes reporting the determined ploidy state to a clinician. In some embodiments, the methods disclosed herein include taking a clinical action, such as performing invasive follow-up testing of chorionic villus sampling or amniocentesis, preparing the trisomic individual for birth, or elective termination of the trisomic fetus.

This application is a continuation of U.S. Utility Application No. 11/603,406, filed November 28, 2006 (U.S. Patent Application Publication No. 20070184467); U.S. Utility Application No. 12/076,348, filed March 17, 2008 (U.S. Patent Application Publication No. 20080243398); and PCT Application No. PCT/US09/52730, filed August 4, 2009 (PCT No. WO/2010/017214; U.S. Utility Model Application No. PCT/US10/050824, filed September 30, 2010 (PCT Publication No. WO/2011/041485); U.S. Utility Model Application No. 13/110,685, filed May 18, 2011; and PCT Application No. PCT/US12/58578, filed October 3, 2012. These patents are incorporated herein by reference in their entireties. Some of the vocabulary used in this application may have its antecedents in these references. Some of the concepts described herein may be better understood in light of the concepts found in these references.

Screening Maternal Blood Containing Free-floating Fetal DNA The methods described herein can be used to aid in the determination of the genotype of a child, fetus, or other target individual in which the targeted genetic material is found in the presence of an amount of other genetic material. In some embodiments, genotype may refer to the ploidy state of one or more chromosomes, one or more disease-linked alleles, or some combination thereof. In this disclosure, the discussion focuses on determining the genetic state of a fetus when fetal DNA is found in maternal blood, but this example is not intended to be limiting to the possible situations in which the method can be applied. Furthermore, the method is applicable when the amount of target DNA is present in any ratio to non-target DNA, for example, the target DNA may constitute anywhere between 0.000001% and 99.999999% of the DNA present. Furthermore, the non-target DNA does not necessarily have to come from one individual, or even from related individuals, so long as genetic data from some or all of the relevant non-target individual(s) is known. In an embodiment, the methods disclosed herein can be used to determine fetal genotype data from maternal blood containing fetal DNA, and can be used when a pregnant woman has multiple fetuses in her uterus, or when other contaminating DNA, such as DNA from other preborn siblings, may be present in the sample.

This technique can exploit the phenomenon of fetal blood cells entering the maternal circulation through the placental villi. Normally, only a very small number of fetal cells enter the maternal circulation in this way (not enough to give a positive Kleihauer-Betke test for feto-maternal hemorrhage). Fetal cells can be selected and analyzed by various techniques to look for specific DNA sequences, but without the inherent risks of invasive procedures. This technique can also exploit the phenomenon of free-floating fetal DNA entering the maternal circulation by DNA release after apoptosis of placental tissue, if the placental tissue in question contains DNA of the same genotype as the fetus. The free-floating DNA found in maternal plasma has been shown to contain as much as 30-40% fetal DNA.

In some embodiments, blood can be drawn from a pregnant woman. Studies have shown that maternal blood can contain small amounts of free-floating DNA from the fetus in addition to free-floating DNA of maternal origin. Furthermore, in addition to many blood cells of maternal origin, there may also be enucleated fetal blood cells that contain DNA of fetal origin, which generally do not contain nuclear DNA. There are many methods known in the art for isolating fetal DNA or making fractions enriched in fetal DNA. For example, chromatography has been shown to make certain fractions enriched in fetal DNA.

Once one has a sample of maternal blood, plasma or other bodily fluid that has been drawn relatively non-invasively and contains a quantity of fetal DNA, either cellular or free-floating, enriched in its ratio to maternal DNA, or in its original ratio, one can determine the genotype of the DNA found in the sample. In some embodiments, blood can be drawn using a needle to withdraw blood from a vein, for example, the basilic vein. The methods described herein can be used to determine fetal genotype data. For example, the methods can be used to determine the ploidy state at one or more chromosomes, and the methods can be used to determine the identity of a SNP or set of SNPs, including insertions, deletions, and translocations. The methods can be used to determine one or more haplotypes, including the parent from which one or more genotypic features originate.

Note that this method works with any nucleic acid that can be used for any genotyping and/or sequencing method, e.g., the ILLUMINA INFINIUM ARRAY platform, AFFYMETRIX GENECHIP, ILLUMINA GENOME ANALYZER, or LIFE TECHNOLOGIES' SOLID SYSTEM. This includes free-floating DNA extracted from plasma or its amplification (e.g., whole genome amplification, PCR); genomic DNA from other cell types (e.g., human lymphocytes from whole blood) or its amplification. To prepare DNA, any extraction or purification method that produces genomic DNA suitable for one of these platforms will work as well. This method may work equally well with samples of RNA. In certain embodiments, the samples may be stored in a way that minimizes degradation (e.g., frozen at about -20°C or lower).

PARENTAL SUPPORT Some embodiments may be used in combination with the PARENTAL SUPPORT™ (PS) Method, embodiments of which are described in U.S. Patent Application No. 11/603,406 (U.S. Patent Application Publication No. 20070184467), U.S. Patent Application No. 12/076,348 (U.S. Patent Application Publication No. 20080243398), U.S. Patent Application No. 13/110,685, PCT Application No. PCT/US09/52730 (PCT Publication No. WO/2010/017214), and PCT Application No. PCT/US10/050824 (PCT Publication No. WO/2011/041485), which are incorporated by reference herein in their entireties. These patents are incorporated herein by reference in their entirety. PARENTAL SUPPORT™ is an informatics-based approach that can be used to analyze genetic data. In some embodiments, the methods disclosed herein can be considered as part of the PARENTAL SUPPORT™ method. In some embodiments, the PARENTAL SUPPORT™ method is a collection of methods that can be used to determine the genetic data of a target individual with high accuracy, the genetic data of one or a small number of cells from that individual, or the genetic data of a mixture of DNA consisting of DNA from the target individual and DNA from one or more other individuals, in particular to determine disease-associated alleles, other alleles of interest, and/or the ploidy state of one or more chromosomes in the target individual. PARENTAL SUPPORT™ can refer to any of these methods. PARENTAL SUPPORT™ is an example of an informatics-based method. An exemplary embodiment of the PARENTAL SUPPORT™ method is illustrated in FIGS.

In the PARENTAL SUPPORT™ method, known parental genetic data, i.e., maternal and/or paternal haplotype and/or diploid genetic data, together with incomplete measurements of the meiotic machinery and target DNA, and possibly knowledge of one or more related individuals, together with population-based crossover frequencies, are used to reconstruct in silico, with a high degree of confidence, the genotype at multiple alleles, and/or the ploidy state of the embryo or any target cell(s), and the target DNA that maps to key loci. The PARENTAL SUPPORT™ method allows the reconstruction of not only poorly measured single nucleotide polymorphisms (SNPs), but also insertions and deletions, as well as SNPs or entire regions of DNA that were not measured at all. Furthermore, the PARENTAL SUPPORT™ method allows both the measurement of multiple disease-linked loci as well as screening for aneuploidy from a single cell. In some embodiments, the PARENTAL SUPPORT™ method can be used to characterize one or more cells from an embryo biopsied during an IVF cycle to determine the genetic status of the one or more cells.

The PARENTAL SUPPORT™ method allows cleaning of noisy genetic data. This can be done by estimating the exact genetic alleles in the target genome (embryo) using the genotypes of related individuals (parents) as a reference. PARENTAL SUPPORT™ may be particularly suitable when only a small amount of genetic material is available (e.g., PGD) and when direct measurement of genotype is inherently noisy due to the limited amount of genetic material. PARENTAL SUPPORT™ may be particularly suitable when only a small portion of the available genetic material is from the target individual (e.g., NPD) and when direct measurement of genotype is inherently noisy due to contaminating DNA signals from another individual. The PARENTAL SUPPORT™ method allows highly accurate regular diploid allele arrangements on embryos to be reconstructed along with the copy numbers of chromosome segments, whereas conventional unordered diploid measurements can be characterized by high rates of allele dropout, drop-in, variable amplification bias and other errors.The method can use both the underlying genetic model and the underlying measurement error model.The genetic model can determine both the allele probability at each SNP and the crossover probability between SNPs.The allele probability can be modeled based on the data obtained from parents at each SNP, and the crossover probability between SNPs can be modeled based on the data obtained from the HapMap database developed by the International HapMap Project. Given the appropriate underlying genetic and measurement error models, maximum a posteriori (MAP) estimation can be used, with modifications to make it computationally efficient, to estimate accurate, ordered allele values at each SNP in the embryo.

The techniques outlined above allow, in some cases, the genotype of an individual to be determined by considering very small amounts of DNA from that individual. This may be DNA from one or a few cells, or it may be from the small amounts of fetal DNA found in maternal blood.

Hypothesis In the context of the present disclosure, a hypothesis refers to a possible genetic state. A hypothesis may refer to a possible ploidy state. A hypothesis may refer to a possible allelic state. A set of hypotheses may refer to a set of possible genetic states, a set of possible allelic states, a set of possible ploidy states, or a combination thereof. In some embodiments, a set of hypotheses can be designed such that one hypothesis from the set corresponds to the actual genetic state of any given individual. In some embodiments, a set of hypotheses can be designed such that every possible genetic state can be described by at least one hypothesis from the set. In some embodiments of the present disclosure, an aspect of the method is to determine which hypothesis corresponds to the actual genetic state of the individual in question.

In another embodiment of the present disclosure, a step includes generating a hypothesis. In some embodiments, the hypothesis may be a copy number hypothesis. In some embodiments, the hypothesis includes a hypothesis regarding which chromosomal segments from each of the related individuals correspond genetically to which segments, if any, in other related individuals. Generating a hypothesis may refer to the act of setting limits on variables such that the entire set of possible genetic states under consideration are encompassed by those variables.

"Copy number hypothesis", also referred to as "ploidy hypothesis" or "ploidy state hypothesis", may refer to a hypothesis regarding the possible ploidy state for a given chromosome copy, chromosome type, or section of a chromosome in a target individual. It may also refer to the ploidy state in two or more chromosome types in an individual. A set of copy number hypotheses may refer to a set of hypotheses, each corresponding to a different possible ploidy state in an individual. A set of hypotheses may be for a set of possible ploidy states, a set of possible parental haplotype contributions, a set of possible fetal DNA percentages in a mixed sample, or a combination thereof.

A normal individual contains one of each chromosome type from each parent. However, due to errors in meiosis and mitosis, an individual may have zero, one, two, or more of a given chromosome type from each parent. In practice, it is rare to find more than two copies of a given chromosome from a parent. In this disclosure, in some embodiments, we only consider the possible hypotheses that zero, one, or two copies of a given chromosome come from a parent, with a trivial extension to consider some possible copies originating from a parent. In some embodiments, for a given chromosome, there are nine possible hypotheses: three possible hypotheses for zero, one, or two chromosomes of maternal origin multiplied by three possible hypotheses for zero, one, or two chromosomes of paternal origin. Let (m, f) refer to the hypothesis where m is the number of a given chromosome inherited by the mother, and f is the number of a given chromosome inherited by the father. Thus, the nine hypotheses are (0,0), (0,1), (0,2), (1,0), (1,1), (1,2), (2,0), (2,1), and (2,2). These can also be written as _H00 , _H01 , _H02 , _H10 , _H12 , _H20 , _H21 , and _H22 . Different hypotheses correspond to different ploidy states. For example, (1,1) refers to a normal disomic chromosome, (2,1) refers to a maternal trisomy, and (0,1) refers to a paternal monosomy. In some embodiments, the case where two chromosomes are inherited from one parent and one chromosome is inherited from the other parent can be further divided into two cases: when the two chromosomes are identical (matching copy error) and when the two chromosomes are homologous but not identical (mismatching copy error). In these embodiments, there are 16 possible hypotheses. It should be understood that other sets of hypotheses, and different numbers of hypotheses, can be used.

In some embodiments of the present disclosure, the ploidy hypothesis refers to a hypothesis regarding which chromosomes from other related individuals correspond to chromosomes found in the genome of the target individual. In some embodiments, key to the method is the fact that related individuals can be predicted to share haplotype blocks, and using measured genetic data from related individuals together with knowledge of which haplotype blocks match between the target and related individuals, it is possible to infer accurate genetic data about the target individual with a higher degree of confidence than using the genetic measurements of the target individual alone. Thus, in some embodiments, the ploidy hypothesis may concern not only the number of chromosomes, but also which chromosomes of related individuals are identical or nearly identical to one or more chromosomes of the target individual.

Once a set of hypotheses is defined, the algorithm may operate on the input genetic data to output a determined statistical probability for each of the hypotheses under consideration. The probabilities of the various hypotheses may be determined by mathematically calculating, using the relevant genetic data as input, a value that is equal in probability to that indicated by one or more of the expert techniques, algorithms, and/or methods described elsewhere in this disclosure for each of the various hypotheses.

Once the probabilities of different hypotheses, as determined by multiple techniques, have been estimated, they can be combined. This may require multiplying the probabilities determined by each technique for each hypothesis. The product of the hypothesis probabilities can be normalized. One ploidy hypothesis refers to one possible ploidy state for a chromosome.

The process of "combining probabilities", also referred to as "combining hypotheses" or combining the results of expert techniques, is a concept that should be familiar to those skilled in the art of linear algebra. One possible way of combining probabilities is as follows: When an expert technique is used to evaluate a set of hypotheses given a set of genetic data, the output of the method is a set of probabilities that are one-to-one associated with each hypothesis in the set of hypotheses. When a set of probabilities determined by a first expert technique, each of which is associated with one of the hypotheses in the set, is combined with a set of probabilities determined by a second expert technique, each of which is associated with the same set of hypotheses, the two sets of probabilities are multiplied. This means that for each hypothesis in the set, the two probabilities associated with that hypothesis, determined by the two expert methods, are multiplied together, and the corresponding product is the output probability. This process can be extended to any number of expert techniques. When only one expert technique is used, the output probability is the same as the input probability. When more than two expert techniques are used, the relevant probabilities can be multiplied simultaneously. The products can be normalized so that the probabilities of the hypotheses in the set of hypotheses sum to 100%.

In some embodiments, if the combined probability for a given hypothesis exceeds the combined probability for any of the other hypotheses, that hypothesis may be considered to be determined to be most likely. In some embodiments, if the normalized probability exceeds a threshold, the hypothesis may be determined to be most likely and a ploidy state or other genetic state may be called. In an embodiment, this may mean that the number and identity of chromosomes associated with that hypothesis may be called as a ploidy state. In an embodiment, this may mean that the identity of alleles associated with that hypothesis may be called as an allele state. In some embodiments, the threshold may be between about 50% and about 80%. In some embodiments, the threshold may be between about 80% and about 90%. In some embodiments, the threshold may be between about 90% and about 95%. In some embodiments, the threshold may be between about 95% and about 99%. In some embodiments, the threshold may be between about 99% and about 99.9%. In some embodiments, the threshold may be greater than about 99.9%.

Parental context Parental context refers to the genetic state of a given allele of each of the two relevant chromosomes for one or both of the two parents of the target. Note that in an embodiment, parental context does not refer to the allelic state of the target, but rather to the allelic state of the parents. The parental context for a given SNP may consist of four base pairs, two paternal and two maternal, which may be the same or different from each other. It is common to write "m ₁ m ₂ | f ₁ f ₂ ", where m ₁ and m ₂ are the genetic states of the given SNP on the two maternal chromosomes, and f ₁ and f ₂ are the genetic states of the given SNP on the two paternal chromosomes. In some embodiments, the parental context can be written as "f ₁ f ₂ | m ₁ m ₂ ". Note that the subscripts "1" and "2" indicate the genotype at a given allele on the first and second chromosomes. Note also that the choice of which chromosome is designated "1" and which is designated "2" is arbitrary.

Note that in this disclosure, A and B are often used to generally indicate the identity of a base pair; A or B can equally well indicate C (cytosine), G (guanine), A (adenine) or T (thymine). For example, if at a given SNP-based allele the mother's genotype is T at that SNP on one chromosome and G at that SNP on the homologous chromosome, and the father's genotype at that allele is G at that SNP on both homologous chromosomes, then the target individual's allele can be said to have the parental context AB|BB, and the allele can also be said to have the parental context AB|AA. Note that in theory, all four possible nucleotides may be present at a given allele, so, for example, the mother may have the genotype AT and the father may have the genotype GC at a given allele. However, empirical data shows that in most cases, only two of the four possible base pairs are observed at a given allele. For example, with a single tandem repeat, it is possible to have more than two, more than four, or even more than ten parental situations. The discussion in this disclosure assumes that only two possible base pairs are observed at a given allele, but the embodiments disclosed herein can be modified to take into account cases where this assumption is not true.

"Parental context" may refer to a set or subset of target SNPs that have the same parental context. For example, if 1000 alleles on a given chromosome of a target individual are measured, context AA|BB may refer to the set of all alleles in the 1,000 allele group where the target's mother's genotype is homozygous and the target's father's genotype is homozygous, but the maternal and paternal genotypes at that locus are not similar. If the parental data is not phased, so AB=BA, there are nine possible parental contexts: AA|AA, AA|AB, AA|BB, AB|AA, AB|AB, AB|BB, BB|AA, BB|AB, and BB|BB. If the parental data are phased, thus AB≠BA, there are 16 different possible parental contexts: AA|AA, AA|AB, AA|BA, AA|BB, AB|AA, AB|AB, AB|BA, AB|BB, BA|AA, BA|AB, BA|BA, BA|BB, BB|AA, BB|AB, BB|BA, and BB|BB. With the exception of some SNPs on the sex chromosomes, every SNP allele on a chromosome has one of these parental contexts. The set of SNPs where the parental context for one parent is heterozygous can be referred to as a heterozygous context.

Use of Parental Status in NPD Non-invasive prenatal testing is an important technique that can be used to determine the genetic status of a fetus non-invasively, for example, from genetic material obtained by drawing blood from the pregnant mother. Blood can be separated, plasma isolated, and then plasma DNA isolated. Size selection can be used to isolate DNA of the appropriate length. DNA can be preferentially enriched in a set of loci. This DNA can then be measured by a number of means, such as hybridizing to a genotyping array and measuring fluorescence, or by sequencing on a high-throughput sequencer.

When using sequencing for fetal ploidy calling in the context of non-invasive prenatal diagnosis, there are a number of ways to use sequence data. The most common way sequence data can be used is simply to count the number of reads that map to a given chromosome. For example, consider that one wishes to determine the ploidy state of chromosome 21 of a fetus. Further consider that 10% of the DNA in the sample is composed of DNA of fetal origin and 90% of DNA of maternal origin. In this case, one looks at the average number of reads for a chromosome that can be predicted to be disomic, e.g., chromosome 3, and compares it to the number of reads on chromosome 21 where the reads are adjusted for the number of base pairs on the chromosome that are part of the unique sequence. If the fetus was euploid, one would expect the amount of DNA per unit of genome to be roughly equal at all locations (subject to stochastic variation). On the other hand, if the fetus was trisomic at chromosome 21, one would expect slightly more DNA per genetic unit from chromosome 21 than elsewhere in the genome. In particular, one would expect about 5% more DNA from chromosome 21 in the mixture. When sequencing is used to measure DNA, it is expected that there will be approximately 5% more uniquely mappable reads from chromosome 21 per unique segment than from other chromosomes. The observation of an amount of DNA from a particular chromosome greater than a certain threshold, when adjusted for the number of uniquely mappable sequences to that chromosome, can be used as the basis for diagnosing aneuploidy. Another method that can be used to detect aneuploidy is similar to the one above, except that it can take into account parental circumstances.

When considering which alleles to target, one can take into account the likelihood that some parental situations may be more informative than others. For example, AA|BB and the symmetric situation BB|AA are the most informative situations because the fetus is known to carry a different allele than the mother. For symmetry reasons, both the AA|BB and BB|AA situations can be referred to as AA|BB. Another set of informative parental situations are AA|AB and BB|AB, because in these cases the fetus has a 50% chance of carrying an allele that the mother does not have. For symmetry reasons, both the AA|AB and BB|AB situations can be referred to as AA|AB. A third set of informative parental situations are AB|AA and AB|BB, because in these cases the fetus carries a known paternal allele that is also present in the maternal genome. For symmetry reasons, the AB|AA and AB|BB situations can be referred to as AB|AA. The fourth parent situation is AB|AB, where the fetus has an unknown allele state, and whatever the allele state is, it is one in which the mother has the same allele. The fifth parent situation is AA|AA, where the mother and father are heterozygous.

Different implementations of the presently disclosed embodiments Disclosed herein are methods for determining the ploidy state of a target individual. The target individual may be a blastomere, an embryo or a fetus. In some embodiments of the present disclosure, the method for determining the ploidy state of one or more chromosomes in a target individual may include any of the steps described herein, and combinations thereof:

In some embodiments, the source of genetic material used in determining the genetic state of the fetus may be fetal cells, such as fetal nucleated red blood cells isolated from maternal blood. The method may include obtaining a blood sample from a pregnant mother. The method may include using visual techniques to isolate fetal red blood cells based on the idea that a particular combination of colors is uniquely associated with nucleated red blood cells, and a similar combination of colors is not associated with any other cells present in the maternal blood. The color combination associated with nucleated red blood cells may include the red color of the hemoglobin around the nucleus, which can be made more distinguishable by staining, and the color of the nuclear material, which can be stained, for example, blue. By isolating cells from the maternal blood and spreading them on a slide, and then identifying points where both red (from hemoglobin) and blue (from nuclear material) are found, it may be possible to identify the location of nucleated red blood cells. These nucleated red blood cells can then be extracted using a micromanipulator, and genotyping and/or sequencing techniques can be used to measure genotypic aspects of the genetic material of these cells.

In certain embodiments, nucleated red blood cells are stained with a dye that fluoresces only in the presence of fetal hemoglobin and not in the presence of maternal hemoglobin, thus eliminating any ambiguity as to whether the nucleated red blood cells are of maternal or fetal origin. Some embodiments of the present disclosure may involve staining or otherwise marking the nuclear material. Some embodiments of the present disclosure may involve specifically marking the fetal nuclear material using an antibody specific for fetal cells.

There are many methods for isolating fetal cells from maternal blood, or for isolating fetal DNA from maternal blood, or for enriching samples of fetal genetic material in the presence of maternal genetic material. Some of these methods are listed here, but this is not intended to be an exhaustive list. For convenience, some suitable techniques are listed here: fluorescently or otherwise tagged antibodies, size exclusion chromatography, magnetically or otherwise labeled affinity tags, epigenetic differences, e.g., differential methylation between maternal and fetal cells at specific alleles, density gradient centrifugation followed by CD45/14 depletion and CD71 positive selection from CD45/14 negative cells, single or double Percoll gradients with different osmolalities, or galactose specific lectin methods.

In some embodiments of the present disclosure, the target individual is a fetus, and different genotype measurements are made on multiple DNA samples from the fetus. In some embodiments of the present disclosure, the fetal DNA sample is from isolated fetal cells, which may be mixed with maternal cells. In some embodiments of the present disclosure, the fetal DNA sample is from free-floating fetal DNA, which may be mixed with free-floating maternal DNA. In some embodiments, the fetal DNA sample can be obtained from maternal plasma or maternal blood, which contains a mixture of maternal and fetal DNA. In some embodiments, fetal DNA may be mixed with maternal DNA in maternal:fetal ratios ranging from 99.9:0.1% to 99:1%; 99:1% to 90:10%; 90:10% to 80:20%; 80:20% to 70:30%; 70:30% to 50:50%; 50:50% to 10:90%; or 10:90% to 1:99%; 1:99% to 0.1:99.9%.

The genetic data of the target individual and/or related individuals can be converted from a molecular state to an electronic state by measuring the appropriate genetic material using tools and/or techniques selected from the group including, but not limited to, genotyping microarrays, and high-throughput sequencing. Some high-throughput sequencing methods include Sanger DNA sequencing, pyrosequencing, the ILLUMINA SOLEXA platform, ILLUMINA's GENOME ANALYZER or APPLIED BIOSYSTEM's 454 sequencing platform, HELICOS' TRUE SINGLE MOLECULE SEQUENCEING platform, HALCYON MOLECULAR's electron microscope sequencing, or any other sequencing method. All of these methods physically convert the genetic data stored in a sample of DNA into a collection of genetic data that is typically stored and processed along the way in a memory device.

Genetic data of the relevant individual may be measured by analyzing material selected from the group including, but not limited to, bulk diploid tissue of the individual, one or more diploid cells from the individual, one or more haploid cells from the individual, one or more blastomeres from the target individual, extracellular genetic material found in the individual, extracellular genetic material from the individual found in maternal blood, cells from the individual found in maternal blood, one or more embryos created from gamete(s) from the relevant individual, one or more blastomeres obtained from such embryos, extracellular genetic material found in the relevant individual, genetic material known to originate from the relevant individual, and combinations thereof.

In some embodiments, a set of at least one ploidy state hypothesis can be generated for each of the chromosome types of interest for the target individual. Each ploidy state hypothesis can refer to one possible ploidy state of a chromosome or chromosome segment of the target individual. The set of hypotheses can include some or all of the possible ploidy states that a chromosome of the target individual can be predicted to have. Some of the possible ploidy states can include nullisomy, monosomy, disomy, uniparental disomy, euploidy, trisomy, concordant trisomy, discordant trisomy, maternal trisomy, paternal trisomy, tetrasomy, balanced (2:2) tetrasomy, unbalanced (3:1) tetrasomy, pentasomy, hexasomy, other aneuploidies, and combinations thereof. Any of these aneuploidy states can be mixed or partial aneuploidies, such as unbalanced translocations, balanced translocations, Robertsonian translocations, recombinations, deletions, insertions, crossovers, and combinations thereof.

In some embodiments, knowledge of the determined ploidy status can be used to make a clinical decision. This knowledge is typically stored in a memory device as a physical array of items and can then be converted into a report. The report can then be executed. For example, the clinical decision can be to terminate the pregnancy, or alternatively, the clinical decision can be to continue the pregnancy. In some embodiments, the clinical decision can involve an intervention designed to reduce the severity of the phenotypic manifestation of the genetic disorder, or a decision to take relevant steps to prepare for a special needs child.

In certain embodiments of the present disclosure, any of the methods described herein can be modified to allow multiple targets to be derived from the same target individual, e.g., multiple blood draws from the same pregnant mother. This can improve the accuracy of the model, since multiple genetic measurements can provide more data from which target genotypes can be determined. In certain embodiments, one set of target genetic data serves as the reported primary data, and the other set of target genetic data serves as data to reconfirm the primary target genetic data. In certain embodiments, multiple sets of genetic data, each measured from genetic material obtained from a target individual, are considered in parallel, and thus both sets of target genetic data serve as aids to determine which sections of parental genetic data measured with high accuracy constitute the fetal genome.

In an embodiment, the method can be used for paternity testing. For example, given SNP-based genotype information from the mother and from a man who may or may not be the genetic father, as well as genotype information measured from a mixed sample, it is possible to determine whether the man's genotype information actually represents the actual genetic father of the gestating fetus. A simple way to do this is to simply test for situations where the mother is AA and the possible father is AB or BB. In these cases, one can expect the father to contribute half the time (AA|AB) or always (AA|BB), respectively. Taking into account the predicted ADO, it is straightforward to determine whether the observed fetal SNPs correlate with the possible father's SNPs.

One embodiment of the present disclosure may be as follows: A pregnant woman wishes to know if her fetus has Down's syndrome and/or cystic fibrosis, and she does not want to give birth to a child with either of these conditions. A physician draws the woman's blood and stains the hemoglobin with one marker to appear distinctly red, and the nuclear material with another marker to appear distinctly blue. Because maternal red blood cells are generally anucleated, but a high percentage of fetal cells are known to contain nuclei, the physician is able to visually isolate a number of nucleated red blood cells by identifying cells that exhibit both red and blue colors. The physician removes these cells from the slide with a micromanipulator and sends them to a laboratory where 10 individual cells are amplified and genotyped. Using genetic measurements, the PARENTAL SUPPORT™ method is able to determine that 6 out of 10 cells are maternal blood cells and 4 out of 10 cells are fetal cells. If the pregnant mother has already had a child, PARENTAL SUPPORT™ can also be used to determine that the fetal cells are distinct from the cells of the born child by making reliable allele calls on the fetal cells and showing that they are not similar to the alleles of the born child. Note that this method is a similar concept to the paternity testing embodiment of the present disclosure. The genetic data measured from the fetal cells can be of very poor quality, containing many allele dropouts due to the difficulty of genotyping single cells. Using the measured fetal DNA together with reliable DNA measurements of the parents, clinicians can use PARENTAL SUPPORT™ to estimate aspects of the fetal genome with high accuracy, thereby converting the genetic data contained in the genetic material from the fetus into a predicted fetal genetic state stored on a computer. Clinicians can determine both the ploidy state of the fetus and the presence or absence of multiple disease-linked genes of interest. The fetus is found to be euploid and not a carrier of cystic fibrosis, and the mother decides to continue the pregnancy.

In one embodiment of the present disclosure, a pregnant mother wishes to determine whether her fetus suffers from any whole chromosomal abnormality. The woman goes to her doctor and provides a sample of her blood, and she and her husband provide a sample of their DNA via cheek swab. Laboratory researchers genotype the parental DNA using the MDA protocol to amplify the parental DNA and ILLUMINA INFINIUM arrays to measure parental genetic data at multiple SNPs. They then spun down the blood, took the plasma, and isolated a sample of free-floating DNA using size-exclusion chromatography. Alternatively, they isolated fetal nucleated red blood cells using one or more fluorescent antibodies, for example, antibodies specific for fetal hemoglobin. They then take the isolated or enriched fetal genetic material and amplify it using a library of 70-mer oligonucleotides appropriately designed such that the two ends of each oligonucleotide correspond to the flanking sequences on either side of the target allele. Upon addition of polymerase, ligase, and appropriate reagents, the oligonucleotides underwent gap-filling circularization, capturing the desired allele. Exonuclease was added, heat inactivated, and the products were used as templates for direct PCR amplification. PCR products were sequenced on an ILLUMINA GENOME ANALYZER. Sequence reads were used as input for the PARENTAL SUPPORT™ method, which then predicted the ploidy state of the fetus.

In another embodiment, a couple, the mother of whom is pregnant and has advanced maternal age, would like to know if the pregnant fetus has Down syndrome, Turner syndrome, Prader-Willi syndrome or some other whole chromosomal abnormality. An obstetrician draws blood from the mother and father. The blood is sent to a laboratory where a technician centrifuges the maternal sample to isolate the plasma and buffy coat. The DNA in the buffy coat and the paternal blood sample are converted by amplification, and the genetic data encoded in the amplified genetic material is further converted from molecularly stored genetic data to electronically stored genetic data by subjecting the genetic material to a high-throughput sequencer to measure the parental genotypes. The plasma sample is preferentially enriched in a set of loci using a 5,000-plex hemi-nested targeted PCR method. A mixture of DNA fragments is prepared into a DNA library suitable for sequencing. The DNA is then sequenced using a high-throughput sequencing method, for example, the ILLUMINA GAIIx GENOME ANALYZER. Sequencing converts the information molecularly encoded in DNA into electronically encoded information in computer hardware. Informatics-based techniques, including embodiments disclosed herein, such as PARENTAL SUPPORT™, can be used to determine the ploidy state of the fetus. This can include the steps of: calculating by computer the probability of allele counts at multiple polymorphic loci from DNA measurements made on the prepared sample; generating by computer multiple ploidy hypotheses, each associated with a different possible ploidy state on the chromosome; constructing by computer a joint distribution model for the expected allele counts at multiple polymorphic loci on the chromosome for each ploidy hypothesis; using the joint distribution model and the allele counts measured on the prepared sample, determining by computer the relative probability of each of the ploidy hypotheses; and calling the ploidy state of the fetus by selecting the ploidy state corresponding to the hypothesis with the greatest probability. It is determined that the fetus has Down's syndrome. The report is printed or sent electronically to the pregnant woman's obstetrician, who communicates the diagnosis to the woman. The woman, her husband, and the doctor sit down to discuss options. The couple decide to terminate the pregnancy based on the knowledge that the fetus suffers from a trisomic condition.

In an embodiment, a company may decide to offer a diagnostic technology designed to detect aneuploidies in a gestating fetus from a maternal blood draw. The product may require the mother to visit her obstetrician, who can draw her blood. The obstetrician may also collect a genetic sample from the father of the fetus. The clinician may isolate plasma from the mother's blood and purify DNA from the plasma. The clinician may also isolate the buffy coat layer from the mother's blood and prepare DNA from the buffy coat. The clinician may also prepare DNA from the father's genetic sample. The clinician may add universal amplification tags to DNA in DNA derived from the plasma sample using molecular biology techniques described in this disclosure. The clinician may amplify the universally tagged DNA. The clinician may preferentially enrich DNA by a number of techniques, including capture by hybridization and targeted PCR. Targeted PCR may involve nesting, hemi-nesting or semi-nesting or any other technique that results in efficient enrichment of plasma-derived DNA. Targeted PCR allows for massive multiplexing, for example with 10,000 primers in one reaction volume, where the primers target SNPs on chromosomes 13, 18, 21, X and loci common to both X and Y, and optionally other chromosomes. Selective enrichment and/or amplification may involve tagging each individual molecule with a different tag, molecular barcode, amplification tag, and/or sequencing tag. The clinician can then sequence the plasma sample and, in some cases, the prepared maternal and/or paternal DNA. The molecular biology steps can be performed entirely or in part by the diagnostic box. The sequence data can be fed to a single computer or to another type of computational platform, such as one that can be found in the "cloud". The computational platform can calculate the number of alleles at the targeted polymorphic loci from the measurements made by the sequencer. The computational platform can generate multiple ploidy hypotheses related to nullisomy, monosomy, disomy, concordant trisomy, and discordant trisomy for each of chromosomes 13, 18, 21, X, and Y. The computational platform can build a joint distribution model for the expected allele counts at the target locus on the chromosome for each of the five chromosomes being examined. The computational platform can determine the probability that each of the ploidy hypotheses is true using the joint distribution model and the allele counts measured on the preferentially enriched DNA from the plasma sample. The computational platform can call the ploidy state of the fetus by selecting the ploidy state corresponding to the appropriate hypothesis with the greatest probability for each of chromosomes 13, 18, 21, X, and Y. A report including the called ploidy state can be generated and can be sent electronically to the obstetrician, displayed on an output device, or a hard copy of the printed report can be delivered to the obstetrician. The obstetrician can inform the patient, and if necessary the father of the fetus, who can then decide which clinical options are acceptable to them and which are most desirable.

In another embodiment, a pregnant woman, hereafter referred to as the "mother", may decide that she would like to know if her fetus(es) carry any genetic abnormalities or other conditions. The mother may wish to ensure that there are no gross abnormalities before deciding to continue the pregnancy. The mother may go to her obstetrician, who may also take a sample of the mother's blood. The obstetrician may also take a genetic sample, such as a buccal swab from the mother's cheek. The obstetrician may also take a genetic sample, e.g., a buccal swab, sperm sample, or blood sample, from the father of the fetus. The obstetrician may send the sample to a clinician. The clinician may enrich for a fraction of free floating fetal DNA in the maternal blood sample. The clinician may enrich for a fraction of enucleated fetal blood cells in the maternal blood sample. The clinician may use various aspects of the methods described herein to determine genetic data for the fetus. The genetic data may include the ploidy state of the fetus and/or the identity of one or more disease-linked alleles in the fetus. A report can be generated summarizing the results of the prenatal testing. The report can be delivered or mailed to a physician who can inform the mother of the genetic status of the fetus. The mother can decide to terminate the pregnancy based on the fact that the fetus has one or more chromosomal or genetic abnormalities or undesirable conditions. The mother can similarly decide to continue the pregnancy based on the fact that the fetus does not have any full chromosomal or genetic abnormalities or any genetic condition of interest.

Another example may involve a pregnant woman who has been artificially inseminated by a sperm donor and is pregnant. The woman wishes to minimize the risk that the fetus she is carrying will have a genetic disease. The woman has her blood drawn by a phlebotomist and using the techniques described in this disclosure, three fetal nucleated red blood cells are isolated and tissue samples are also taken from the mother and genetic father. The genetic material from the fetus and the mother and father are optionally amplified and genotyped using an ILLUMINA INFINIUM BEADARRAY and the methods described herein clean the parental and fetal genotypes, phase with high accuracy, and make a ploidy call on the fetus. The fetus is found to be euploid and phenotypic susceptibility is predicted from the reconstructed fetal genotype and a report is generated and sent to the mother's physician so they can determine what clinical decisions may be best.

In an embodiment, the raw genetic material of the mother and father is converted by amplification into a quantity of DNA of similar sequence but greater quantity. The genotyping method then converts the genotype data encoded by the nucleic acid into genetic measurements that can be physically and/or electronically stored in a memory device such as those described above. The relevant algorithms that make up the PARENTAL SUPPORT™ algorithm are translated into a computer program using a programming language, the relevant portions of which are discussed in detail herein. The raw measurement data is then converted into a pattern that indicates a high confidence determination of the ploidy status of the fetus by running the computer program on computer hardware that is organized into a pattern that indicates the raw measurement data, rather than physically coded bits and bytes. The details of this conversion depend on the data itself and the computer language and hardware system used to carry out the methods described herein. The data, physically configured to indicate a high quality fetal ploidy determination, is then converted into a report that can be sent to a health care practitioner. This conversion can be done using a printer or a computer display. The report can be a copy printed on paper or other suitable medium, or the report can be electronic. In the case of electronic reports, they can be transmitted and physically stored on a memory device located on a computer available to the healthcare practitioner. Electronic reports can also be displayed on a screen so that they can be read. In the case of a screen display, the data can be converted into a readable form by causing a physical transformation of pixels on the display device. The transformation can be achieved by physically firing electrons at a phosphorescent screen, by changing an electrical charge that physically changes the transparency of a particular set of pixels on the screen that can be placed in front of a substrate that emits or absorbs photons. This transformation can be achieved by changing the orientation of nanoscale molecules in a liquid crystal, for example, from a nematic phase to a cholesteric or smectic phase, in a particular set of pixels. This transformation can be achieved by an electric current that causes the emission of photons from a particular set of pixels that are made up of a number of light emitting diodes arranged in a meaningful pattern. This transformation can be achieved by any other method used to display information, such as a computer screen or some other output device or method of conveying information. The healthcare practitioner can then act on the report to convert the data in the report into a measure. The action may be to continue or terminate the pregnancy, in which case the pregnant fetus with the genetic abnormality is transformed into a non-viable fetus. The transformations enumerated herein may, for example, be summed up so that the genetic material of the pregnant mother and father can be transformed through a number of steps outlined in this disclosure into a medical decision consisting of aborting the fetus with the genetic abnormality or continuing the pregnancy. Alternatively, a collection of genotypic measurements may be transformed into a report that helps a physician treat a pregnant patient.

In certain embodiments of the present disclosure, the methods described herein can be used to determine the ploidy state of a fetus even when the host mother, i.e., the pregnant woman, is not the biological mother of the fetus she is carrying. In certain embodiments of the present disclosure, the methods described herein can be used to determine the ploidy state of a fetus using only a maternal blood sample, without the need for a paternal genetic sample.

Some of the mathematics in the embodiments disclosed herein make hypotheses regarding a limited number of aneuploidy states. In some cases, for example, only zero, one or two chromosomes are predicted to originate from each parent. In some embodiments of the present disclosure, the mathematical derivation can be expanded to take into account other forms of aneuploidy, such as quadrosomies, pentasomies, and hexasomies, where three chromosomes originate from one parent, without changing the basic concepts of the present disclosure. At the same time, it is possible to focus only on a smaller number of ploidy states, such as trisomies and disomies. It should be noted that ploidy determinations that indicate a non-integer number of chromosomes may indicate mosaicism in the sample of genetic material.

In some embodiments, the genetic abnormality is a type of aneuploidy, such as Down's syndrome (or trisomy 21), Edwards syndrome (trisomy 18), Patau syndrome (trisomy 13), Turner syndrome (45X), Klinefelter syndrome (males with two X chromosomes), Prader-Willi syndrome, and DiGeorge syndrome (UPD15). Congenital disorders such as those listed in the previous sentence are generally undesirable, and the knowledge that a fetus suffers from one or more phenotypic abnormalities can provide the basis for deciding to terminate the pregnancy, take necessary measures to prepare for the birth of a special needs child, or take some therapeutic approach intended to reduce the severity of the chromosomal abnormality.

In some embodiments, the methods described herein can be used very early in pregnancy, for example, as early as 4 weeks, as early as 5 weeks, as early as 6 weeks, as early as 7 weeks, as early as 8 weeks, as early as 9 weeks, as early as 10 weeks, as early as 11 weeks, and as early as 12 weeks.

In some embodiments, the methods disclosed herein are used in the context of preimplantation genetic diagnosis (PGD) for selecting embryos during in vitro fertilization, where the target individual is an embryo, and parental genotype data can be used to make ploidy determinations for the embryo from single or two cell biopsies from day 3 embryos or sequencing data from trophectoderm biopsies of day 5 or 6 embryos. In the PGD setting, only the DNA of the offspring is measured and only a small number of cells are examined, typically 1-5, but at most 10, 20 or 50. The total number of starting copies of alleles A and B (at the SNP) is then trivially determined by the offspring's genotype and number of cells. In NPD, the number of starting copies is very high, and therefore the allele ratios after PCR are expected to accurately reflect the starting ratios. However, the small number of starting copies in PGD means that contamination and imperfect PCR efficiency have a non-trivial effect on the allele ratios after PCR. This effect may be more important than read depth in predicting the variance in measured allele ratios after sequencing. The distribution of measured allele ratios given known offspring genotypes can be generated by Monte Carlo simulation of the PCR process based on PCR probe efficiency and contamination probability. Given the distribution of allele ratios for each of the possible offspring genotypes, the likelihood of the various hypotheses can be calculated similarly as described for NIPD.

Maximum Likelihood Estimation The majority of methods known in the art for detecting the presence or absence of a biological phenomenon or medical condition involve using a single hypothesis rejection test, measuring a metric that correlates with the condition, and if the metric is on one side of a given threshold, the condition is present, and if the metric is on the other side of the threshold, the condition is not present. In a single hypothesis rejection test, we simply look at the null distribution when making a decision between the null hypothesis and the alternative hypothesis. Without taking into account the alternative distribution, we cannot estimate the likelihood of each hypothesis given the observed data, and therefore cannot calculate the confidence in the call. Thus, we use a single hypothesis rejection test to get a yes or no answer without sensitivity to the confidence associated with a particular case.

In some embodiments, the methods disclosed herein can detect the presence or absence of a biological phenomenon or medical condition using maximum likelihood methods. Maximum likelihood methods are a substantial improvement over methods using single hypothesis rejection, since the threshold for calling the absence or presence of a condition can be appropriately adjusted for each case. This is particularly relevant for diagnostic techniques that aim to determine the presence or absence of aneuploidy in a gestating fetus from genetic data available from a mixture of fetal and maternal DNA present in the free-floating DNA found in maternal plasma. This is due to the fact that the fraction of fetal DNA in the plasma introduces a fractional change in the optimal threshold for calling aneuploidy versus euploidy. As the fetal fraction drops, the distribution of data associated with aneuploidy becomes more and more similar to the distribution of data associated with euploidy.

Maximum likelihood estimation uses the distributions associated with each hypothesis to estimate the likelihood of the data conditioned on each hypothesis. These conditional probabilities can then be converted into hypothesis calls and confidences. Similarly, maximum a posteriori estimation uses the same conditional probabilities as maximum likelihood estimation, but also incorporates prior populations in selecting the best hypothesis and determining confidence.

Thus, using maximum likelihood estimation (MLE) or the closely related maximum a posteriori (MAP) technique provides two advantages: it increases the likelihood of a correct call and also allows a confidence to be calculated for each call. In an embodiment, the step of selecting the ploidy state corresponding to the hypothesis with the greatest probability is performed using maximum likelihood estimation or maximum a posteriori estimation. In an embodiment, a method is disclosed for determining the ploidy state of a fetus during pregnancy, which involves taking any method currently known in the art that uses single hypothesis rejection and reformulating it to use MLE or MAP techniques. Some examples of methods that can be significantly improved by applying these techniques can be found in U.S. Pat. No. 8,008,018, U.S. Pat. No. 7,888,017 or U.S. Pat. No. 7,332,277.

In an embodiment, a method is described for determining the presence or absence of fetal aneuploidy in a maternal plasma sample containing fetal genomic DNA and maternal genomic DNA, comprising obtaining a maternal plasma sample, measuring DNA fragments found in the plasma sample using a high-throughput sequencer, mapping sequences to chromosomes and determining the number of sequence reads mapped to each chromosome, calculating the percentage of fetal DNA in the plasma sample, calculating a predicted distribution of the amount of the target chromosome predicted to be present if the second target chromosome is euploid, and one or more predicted distributions predicted if the chromosome is aneuploid, using the fetal fraction and the number of sequence reads mapped to one or more reference chromosomes predicted to be euploid, and using MLE or MAP to determine which distribution is most likely to be accurate, thereby indicating the presence or absence of fetal aneuploidy. In an embodiment, measuring the DNA from the plasma may include performing massively parallel shotgun sequencing. In an embodiment, measuring DNA from the plasma sample may include sequencing DNA that is preferentially enriched at multiple polymorphic or non-polymorphic loci, for example by targeted amplification. The multiple loci can be designed to target one or a few suspected aneuploidy chromosomes and one or a few reference chromosomes. The purpose of preferential enrichment is to increase the number of informative sequence reads for determining ploidy.

Informatics Methods of Ploidy Calling Described herein are methods for determining the ploidy status of a fetus taking into account sequence data. In some embodiments, the sequence data can be measured on a high-throughput sequencer. In some embodiments, the sequence data can be measured on DNA originating from free-floating DNA isolated from maternal blood, where the free-floating DNA includes some DNA of maternal origin and some DNA of fetal/placental origin. In this section, an embodiment of the present disclosure is described in which the proportion of fetal DNA in the analyzed mixture is unknown and is assumed to be estimated from the data. The proportion of fetal DNA in the mixture ("fetal fraction") or the percentage of fetal DNA can be measured by other methods, and embodiments are also described in which it is assumed to be known in determining the ploidy status of the fetus. In some embodiments, the fetal fraction can be calculated using only genotyping measurements made on the maternal blood sample itself, which is a mixture of fetal and maternal DNA. In some embodiments, the percentage can also be calculated using the measured or otherwise known genotype of the mother and/or the measured or otherwise known genotype of the father. In another embodiment, the ploidy state of the fetus can be determined simply based on the percentage of fetal DNA calculated for the chromosome in question compared to the percentage of fetal DNA calculated for a reference chromosome that is assumed to be disomic.

In a preferred embodiment, when N SNPs are observed and analyzed for a particular chromosome:
- A set of NR floating DNA sequence measurements S = ( _s1 ,..., _sNR ). Since this method utilizes SNP measurements, all sequence data corresponding to non-polymorphic loci can be ignored. In simplified form, if counts (A,B) are obtained for each SNP, where A and B correspond to the two alleles present at a given locus, then S can be written as S = (( _a1 , _b1 ),...,( _aN , _bN )), where _ai are the A counts on SNP i and _bi are the B counts on SNP i, Σi _{= 1:N} ( _ai + _bi ) = NR, and - The parental data is
o Genotypes from SNP microarray or other intensity-based genotyping platform: mother M = ( _m1 ,..., _mN ), father F = ( _f1 ,..., _fN ), _mi , _fi ∈ (AA,AB,BB), and/or o Sequence data measurements: composed of NRM maternal measurements SM = ( _sm1 ,..., _smnrm ), NRF paternal measurements SF = ( _sf1 ,..., _sfnrf ). Similar to the above simplification, if counts (A,B) are available for each SNP, then SM = (( _am1 , _bm1 ),...,( _amN , _bmN )), SF = (( _af1 , _bf1 ),...,( _afN , _bfN )).

Collectively, the maternal, paternal and child data are denoted as D = (M, F, SM, SF, S). Note that while obtaining parental data is desirable and increases the accuracy of the algorithm, parental data, especially paternal data, is not necessary. This means that it is possible to obtain very accurate copy number results even in the absence of maternal and/or paternal data.

The best copy number estimate (H ^†* ) can be derived by maximizing the log-likelihood of the data, LIK(D|H), over all possible hypotheses (H). In particular, the relative probability of each of the ploidy hypotheses can be determined using a joint distribution model and the number of alleles measured in the prepared samples, and the hypothesis most likely to be correct can be determined as follows:
Similarly the posterior hypothesis likelihood, given the data, is:
where priorprob(H) is the prior probability assigned to each hypothesis H based on the model design and prior knowledge.
It is also possible to use prior probabilities to obtain maximum a posteriori estimates:

In an embodiment, copy number hypotheses that can be taken into account are:
Monosomy:
Maternal H10 (one copy from the mother)
Paternal H01 (one copy from the father)
Disomy: H11 (one copy from each parent)
Simple trisomies, crossovers not considered:
Maternal: H21_concordant (two identical copies from mother, one copy from father), H21_discordant (both copies from mother, one copy from father)
Paternal: H12_concordant (one copy from mother, two identical copies from father), H12_discordant (one copy from mother, both copies from father)
- Consider complex trisomies and crossovers (using a joint distribution model):
Maternal H21 (two copies from the mother, one copy from the father),
Paternal H12 (one copy from the mother, two copies from the father)

In other embodiments, other ploidy states can be considered, such as nullisomy (H00), uniparental disomy (H20 and H02), and tetrasomy (H04, H13, H22, H31 and H40).

In the absence of crossovers, each trisomy, whether of mitosis, meiosis I or meiosis II origin, is either a concordant or discordant trisomy. Due to crossovers, the true trisomy is usually a combination of the two. First, a method is described to derive the likelihood of the hypotheses for simple hypotheses. Second, a method is described to derive the likelihood of the hypotheses for compound hypotheses that combine the individual SNP likelihoods with crossovers.

LIK(D|H) for a simple hypothesis
In an embodiment, for a simple hypothesis, LIK(D|H) can be determined as follows: For a simple hypothesis H, the log-likelihood LIK(H) of hypothesis H for the entire chromosome can be calculated as the sum of the log-likelihoods of the individual SNPs, given the known or derived child fraction cf. In an embodiment, cf can be derived from the data.
This hypothesis does not assume linkage between any SNPs and therefore does not utilize a joint distribution model.

In some embodiments, the log-likelihood can be determined for each SNP. For a particular SNP i, given a hypothesis H about fetal ploidy and percent fetal DNA cf, the log-likelihood of the observed data D is:
where m is the possible true genotype of the mother, f is the possible true genotype of the father, m, f∈{AA,AB,BB}, and c is the possible child genotype given hypothesis H. In particular, for monosomy, c∈{A,B}, for disomy, c∈{AA,AB,BB}, and for trisomy, c∈{AAA,AAB,ABB,BBB}.

Genotype prior frequency: p(m|i) is the general prior probability of the maternal genotype m at SNP i based on the known population frequency at SNP I, denoted as pAi. Specifically,
p(AA|pA _i )=(pA _i ) ² , p(AB|pA _i )=2(pA _i )*(1-pA _i ), p(BB|pA _i )=(1-pA _i ) ²
The probability of the paternal genotype, p(f|i), can be determined similarly.

The probability of true child: p(c|m,f,H) is the probability of obtaining true child genotype=c given parents m, f, and the assumed hypothesis H, which can be easily calculated. For example, p(c|m,f,H) for H11, H21 match, and H21 mismatch are shown below:

Data likelihood: P(D|m,f,c,H,i,cf) is the probability of a given data D at SNP i given the true maternal genotype m, the true paternal genotype f, the true child genotype c, the hypothesis H and the child proportion cf. This can be decomposed into the probabilities of the maternal, paternal and child data as follows:
P(D|m,f,c,H,cf,i)=P(S|m,i)P(M|m,i)P(S|f,i)P(F|f,i)P(S|m,c,H,cf,i)

Likelihood of maternal SNP array data: Assuming the SNP array genotype is accurate, the probability _m of maternal SNP array genotype data at SNP i compared to the true genotype m is simply
It is.

Likelihood of maternal sequence data: The probability of maternal sequence data at SNP i, for count S _i = (am _i , bm _i ), is a binomial probability defined without extra noise or bias as P(SM|m,i) = P _X|m (am _i ), where X|m ∼ Binom(p _m (A), am _i + bm _i ), and p _m (A) has the values shown in the table below.

Likelihood of paternal data: Similar formulas apply to the likelihood of paternal data. Note that it is possible to determine the child's genotype without using parental data, especially the paternal data. For example, if the paternal genotype data F is not available, one can simply use P(F|f,i) = 1. If the paternal sequence data SF is not available, one can simply use P(SF|f,i) = 1.

In some embodiments, the method involves building a joint distribution model for the expected allele counts at multiple polymorphic loci on a chromosome for each ploidy hypothesis, and one method for achieving such an objective is described herein. Likelihood of free fetal DNA data: P(S|m,c,H,cf,i) is the probability of free fetal DNA sequence data at SNP i, given the true maternal genotype m, the true child genotype c, the child copy number hypothesis H, and assuming child proportion cf. This is in fact the probability of sequence data S at SNP I, P(S|m,c,H,cf,i)=P(S|μ(m,c,cf,H),i), given the true probability of A content at SNP i, μ(m,c,cf,H).
For counts, S _i =(a _i ,b _i ), with no extra noise or bias in the data,
P(S|μ(m,c,cf,H),i)= _Px (a _i )
where X ∼ Binom(p(A), a _i + b _i ) and p(A) = μ(m, c, cf, H). In the more complicated case where the exact alignment and (A, B) counts per SNP are unknown, P(S|μ(m, c, cf, H), i) is a combination of integrated binomials.

Probability of true A content: The true probability μ(m,c,cf,H) of A content at SNP i in this mother/child mixture, given true maternal genotype=m, true child genotype=c, and overall child proportion=cf, is
where #A(g) = number of As in genotype g, n _m = 2 is the mother's somy, and n _c is the ploidy of the child under hypothesis H (1 is monosomy, 2 is disomy, and 3 is trisomy).

Using the Joint Distribution Model: LIK(D|H) for Multiple Hypotheses
In some embodiments, the method involves building a joint distribution model of expected allele numbers at multiple polymorphic loci on a chromosome for each ploidy hypothesis, and one method to achieve such an objective is described herein. In many cases, trisomies are not purely concordant or discordant, usually due to crossovers, and therefore in this section results are derived for the composite hypotheses H21 (maternal trisomy) and H12 (paternal trisomy), which combine concordant and discordant trisomies, taking into account possible crossovers.

In the case of trisomy, in the absence of crossover, the trisomy is simply a concordant or discordant trisomy. A concordant trisomy is when a child inherits two copies of the same chromosomal segment from one parent. A discordant trisomy is when a child inherits one copy of each homologous chromosomal segment from a parent. With crossover, some segments of a chromosome may have concordant trisomy and other parts may have discordant trisomy. This section describes how to build a joint distribution model of the expected number of alleles at several loci for a set of alleles, for heterozygosity rates, i.e., for one or more hypotheses.

Assume that for SNP i, LIK(D|Hm,i) is the fit to the concordance hypothesis _Hm , LIK(D|Hu,i) is the fit to the mismatch hypothesis _Hu , and pc(i) = the probability of crossover between SNP i-1 and SNP i. Then the complete likelihood can be calculated as follows:
where LIK(D|E,1:N) is the final likelihood of hypothesis E for SNP 1:N. E = final SNP hypothesis, E ∈ (Hm,Hu). Recursively, we can calculate:
where ∼E is a hypothesis other than E (non-E), and the hypotheses considered are H _m and H _u . In particular, the likelihood of 1:i SNP can be calculated based on the likelihood of SNPs 1 to (i-1) multiplied by the likelihood of SNP i with either the same hypothesis without crossover or the opposite hypothesis with crossover:
For SNP 1, i=1, LIK(D|E,1:1)=LIK(D|E,1),
For SNP 2, i=2, LIK(D|E,1:2)=LIK(D|E,2)+log(exp(LIK(D|E,1))*(1-pc(2))+exp(LIK(D|~E,1))*pc(2));
The same is true for i=3:N.

In some embodiments, the fraction of offspring can be determined. The fraction of offspring may refer to the proportion of sequences originating from the offspring in a mixture of DNA. In the context of non-invasive prenatal testing, the fraction of offspring may refer to the proportion of sequences in maternal plasma originating from the fetus or from portions of the placenta with fetal genotype. The fraction of offspring may refer to the fraction of offspring in a sample of DNA prepared from maternal plasma, which may be enriched for fetal DNA. One purpose of determining the fraction of offspring in a sample of DNA is for use in an algorithm that can make ploidy calls on the fetus, and thus the fraction of offspring may refer to any sample of DNA analyzed by sequencing for non-invasive prenatal testing.

Some of the algorithms presented in this disclosure that are part of the method for non-invasive prenatal aneuploidy diagnosis assume a known offspring proportion, but this is not always the case. In one embodiment, it is possible to find the most likely offspring proportion by maximizing the likelihood of disomy on a selected chromosome, with or without the presence of parental data.

In particular, assume LIK(D|H11,cf,chr) = log likelihood as above for the disomy hypothesis and for the proportion of offspring on chromosome chr,cf. The full likelihood for the selected chromosomes in Cset (usually 1:16), assuming euploidy, is:
It is.
The proportion of most likely offspring (cf ^†* ) is derived as cf ^* = argmax _cf LIK(cf).

Any set of chromosomes can be used. It is also possible to derive offspring fractions without assuming euploidy in the reference chromosomes. Using this method, it is possible to determine offspring fractions for any of the following situations: (1) having array data on the parents and shotgun sequencing data on maternal plasma; (2) having array data on the parents and targeted sequencing data on maternal plasma; (3) having targeted sequencing data on both the parents and maternal plasma; (4) having targeted sequencing data on both the mother and the maternal plasma fraction; (5) having targeted sequencing data on the maternal plasma fraction; (6) other combinations of parental and offspring fraction measurements.

In some embodiments, data dropout can be incorporated by informatics methods. This can result in more accurate ploidy determination results. Elsewhere in this disclosure, it has been assumed that the probability of A occurring is a linear function of the true maternal genotype, the true child genotype, the proportion of children in the mixture, and the child copy number. For example, instead of measuring the true child AB in the mixture, it is possible that the mother or child alleles will drop out, which may occur if only sequences that map to allele A are measured. The parent dropout rate for the genomic Illumina data can be denoted as d _pg , the parent dropout rate for the sequence data as d _ps , and the child dropout rate for the sequence data as d _cs . In some embodiments, it can be assumed that the maternal dropout rate is 0 and the child dropout rate is relatively low, in which case the results are not significantly affected by the dropout. In some embodiments, the probability of allele dropout may be large enough that it significantly impacts the predicted ploidy call. In such cases, allele dropout is now incorporated into the algorithm:

Parental SNP array data dropout: For the maternal genome data M, let the genotype after dropout be m _d .
In the formula, as above,
and P(m _d |m) is the likelihood of a possible dropout genotype m _d given the true genotype m for a dropout rate d, defined as follows:
A similar formula applies to the paternal SNP array data.

Parent sequence data dropout: For the maternal sequence data SM,
where P( _md |m) is as defined in the previous section and the Px _|md ( _ami ) probability from the binomial distribution is defined above in the Likelihood of Parental Data section. A similar formula can be applied to the paternal sequence data.

Floating DNA sequence data dropout:
where P(S|μ(m _d , _cd , cf, H),i) is as defined in the section on likelihood of floating data.

In one embodiment, p(m _d |m) is the probability of observed maternal genotype m _d given the true maternal genotype m and assuming a dropout rate d _ps , and p(c _d |c) is the probability of observed child genotype c _d given the true child genotype c and assuming a dropout rate d _cs . Let nA _T = the number of A alleles in true genotype c, nA _D = the number of A alleles in observed genotype c _d , nA _T ≧nA _D , and similarly, nB _T = the number of B alleles in true genotype c, nB _D = the number of B alleles in observed genotype c _d , nB _T ≧nB _D , and d = the dropout rate.
It is.

In one embodiment, informatics methods may incorporate random and consistent biases. Ideally, there is no consistent sampling bias or random noise (in addition to binomial variation) per SNP in the sequence counts. In particular, for maternal genotype m, true offspring genotype c and offspring proportion cf at SNP i, and X=number of A in the set of (A+B) reads at SNP i, X acts as X~Binominal(p,A+B), where p=μ(m,c,cf,H)=true probability of A content.

In an embodiment, a random bias may be incorporated using informatics methods. In most cases, one assumes that the measurements are biased, and therefore the probability of A occurring at this SNP is equal to q, which is slightly different from p, as defined above. How different p and q are depends on the accuracy of the measurement process and a number of other factors, and can be quantified by the standard deviation of q away from p. In an embodiment, q can be modeled as having a beta distribution, with parameters α, β depending on the mean of that distribution centered on p, and some particular standard deviation s. In particular, this gives X|q ∼ Bin(q,D _i ), where q ∼ Beta(α,β). If E(q)=p, V(q)=s ² , then the parameters α, β can be derived as α=pN, β=(1−p)N,
It is.

This is the definition of beta binomial distribution, which samples from a binomial distribution with a variability parameter q, which follows a beta distribution with mean p. Thus, in an unbiased setup, at SNP i, assuming the true maternal genotype (m), the parental sequence data (SM) probability given the maternal sequence A counts on SNP i (am _i ) and the maternal sequence B counts on SNP i (bm _i ) can be calculated as follows: P(SM|m,i)=PX _|m (am _i ), where X|m∼Binom( _pm (A),am _i +bm _i ).

Now, including a random bias with standard deviation s, we get:
X|m ∼ BetaBinom(p _m (A), am _i + bm _i , s).

In the absence of bias, assuming the true maternal genotype (m), the true child genotype (c), the child proportion (cf), and assuming the child hypothesis H, the probability of the maternal plasma DNA sequence data (S) taking into account the floating DNA sequence A count (ai) on SNP i and the floating sequence B count (bi) on SNP i can be calculated as follows:
P(S|m, c, cf, H, i)=Px(a _i )
where X~Binom(p(A), _ai + _bi ), p(A) = μ(m, c, cf, H).

In one embodiment, including a random bias with standard deviation s makes this X~BetaBinom(p(A), _ai + _bi , s), where the amount of extra variation is specified by the deviation parameter s or equivalently N. Smaller values of s (or larger values of N) make the distribution closer to a standard binomial distribution. To estimate the amount of bias, i.e., N above, from the obvious situations AA|AA, BB|BB, AA|BB, BB|AA, we can estimate the above probabilities:
Depending on the behavior of the data, N can be made to be a constant regardless of the read depth a _i + b _i or a _i + b _i function, resulting in less bias for larger read depths.

In an embodiment, an informatics method may incorporate a consistent SNP-by-SNP bias. Due to artifacts in the sequencing process, some SNPs may have consistently low or high counts regardless of their true A content. Assume that SNP i consistently adds a bias of w _i percent to the A counts. In some embodiments, this bias can be estimated from a set of training data derived under the same conditions and added back to the parent sequence data estimates as follows:
P(SM|m,i)=PX _|m ( _ami )where X|m~BetaBinom( _pm (A)+ _wi , _ami + _bmi ,s)
And the floating DNA sequence data probability estimate is:
P(S|m,c,cf,H,i)= _PX ( _ai )where X~BetaBinom(p(A)+ _wi , _ai + _bi ,s)
It becomes.

In some embodiments, the method can be written to specifically take into account additional noise, differential sample quality, differential SNP quality, and random sampling bias. An example of this is shown here. This method has been shown to be particularly useful in situations where data were generated using a massively multiplexed mini-PCR protocol, and was used in Experiments 7-13. The method involves several steps, each of which introduces a different type of noise and/or bias into the final model:
(1) Assume that a first sample containing a mixture of maternal and fetal DNA contains an amount of original DNA of size = N ₀ molecules, typically in the range of 1,000-40,000, p = true %ref.
(2) In amplification using universal ligation adaptors, we assume that N ₁ molecules are sampled; typically, N ₁ -N ₀ /2 molecules and random sampling bias is introduced due to sampling. An amplified sample may contain any number of molecules N ₂ , with N ₂ >>N _1. Let X ₁ denote the amount of reference locus (per SNP) among the N ₁ sampled molecules, with a variation of p ₁ =X ₁ /N ₁ that introduces random sampling bias throughout the rest of the protocol. We include this sampling bias in the model by using a beta-binomial (BB) distribution instead of using a simple binomial distribution model. The parameter N of the beta-binomial distribution can later be estimated from the training data, for each sample, after adjusting for leakage and amplification bias for SNPs with 0<p<1. Leakage is the tendency of a SNP to be read incorrectly.
(3) The amplification step amplifies any allele bias, thus introducing amplification bias due to possible uneven amplification. Assuming that one allele at a locus is amplified f-fold and the other allele at that locus is amplified g-fold, then f=ge ^b , with b=0 indicating no bias. The bias parameter b centers around 0 and indicates how much more or less allele A is amplified at a particular SNP as opposed to allele B. The parameter b may vary from SNP to SNP. The bias parameter b can be estimated for each SNP, for example, from training data.
(4) The sequencing step involves sequencing a sample of the amplified molecules. In this step, leakage may exist, which is a situation where a SNP is read incorrectly. Leakage may result from a variety of issues, resulting in a SNP not being read as the correct allele A, but as another allele B found at the locus, or as alleles C or D, which are not typically found at the locus. Assume that sequencing measures sequence data for a number of DNA molecules from an amplified sample of size _N3 , _N3 < _N2 . In some embodiments, _N3 may be in the range of 20,000-100,000; 100,000-500,000; 500,000-4,000,000; 4,000,000-20,000,000; or 20,000,000-100,000,000. Each sampled molecule has a probability _pg of being read correctly, in which case it is denoted correctly as allele A. A sample may be read incorrectly as an allele unrelated to the original molecule with probability 1-pg, likely allele A with probability pr, likely allele B with probability _pm , or likely allele C or allele D with probability _p0 , where _pr + _pm + _p0 =1. The parameters _pg , _pr , _pm , and _p0 are estimated for each SNP from the training data.

Different protocols involve similar steps and may involve variations in the molecular biology steps resulting in different amounts of random sampling, different levels of amplification and different leakage biases. The following model can be applied equally well to each of these cases. The model for the amount of sampled DNA is given for each SNP by:
_X3 ~ BetaBinomial(L(F(p,b), _pr , _pg ),N*H(p,b))
where p = true amount of reference DNA, b = bias per SNP, p _g is the probability of a correct read as above, and p _r is the probability of a read that is incorrectly read but appears to be the correct allele by chance, in the case of a bad read, as above:
F(p,b) = pe ^b / (pe ^b + (1-p)), H(p,b) = ( ^eb p + (1-p)) ² / ^eb , and L(p, _pr , _pg ) = p * _pg + _pr * (1- _pg ).

In some embodiments, the method uses a beta binomial distribution instead of a simple binomial distribution. This addresses random sampling bias. The beta binomial parameter N is estimated for each sample as needed. Amplification bias is addressed using bias corrections F(p,b), H(p,b) instead of simply p. The bias parameter b is estimated from the training data in advance for each SNP.

In some embodiments, the method uses a leakage correction L(p, _pr , _pg ) instead of simply p, which accounts for bias due to leakage, i.e., SNP and sample quality variations. In some embodiments, the parameters _pg , _pr , _po are estimated from training data in advance for each SNP. In some embodiments, the parameters _pg , _pr , _po can be updated using the current sample run to account for sample quality variations.

The models described herein are fairly general and can account for both differential sample quality and differential SNP quality. Different samples and SNPs are treated differently, as illustrated by the fact that some embodiments use a beta binomial distribution whose mean and variance are a function of the original DNA amount, as well as sample and SNP quality.

Modeling the Platform Consider a single SNP with a predicted allele ratio r present in plasma (based on maternal and fetal genotypes). The predicted allele ratio is defined as the predicted proportion of allele A in the combination of maternal and fetal DNA. For a maternal genotype _gm and a child's genotype _gc , the predicted allele ratio is given by Equation 1, assuming the genotypes are similarly expressed in allele ratios.
r = fg _c + (1 - f) g _m (1)

An observation at a SNP consists of the number of reads mapped at each allele present, n _a and n _b , summing to a read depth d. We assume that a threshold has already been applied to the mapping probability and phred score, so that the mapping and allele observations can be considered accurate. The phred score is a numerical measure of the probability that a particular measurement at a particular base is incorrect. In an embodiment, if the base is measured by sequencing, the phred score can be calculated from the ratio of the intensity of the dye corresponding to the called base to the intensity of the dye of the other base. The simplest model for observing the likelihood is a binomial distribution, which assumes that each of the d reads is independently drawn from a large pool with an allele ratio r. Equation 2 describes this model.

The binomial model can be extended in a number of ways. If the maternal and fetal genotypes are either all A or all B, the expected allele ratios in plasma will be 0 or 1, and the binomial probabilities are not well-defined. In practice, unexpected alleles are sometimes observed in practice. In one embodiment, the corrected allele ratios
Using the above method, it is possible to make the unexpected alleles minor. In an embodiment, it is possible to use the training data to model the ratio of unexpected alleles appearing on each SNP, and to use this model to correct the expected allele ratio. If the expected allele ratio is not 0 or 1, the observed allele ratio may not converge to the expected allele ratio at a high enough read depth due to amplification bias or other phenomena. The allele ratio can then be modeled as a beta distribution centered on the expected allele ratio, resulting in a beta binomial distribution for P(n _a , n _b |r) with a larger variance than the binomial distribution.

The platform model for response at a single SNP is defined as F(a, b, _gc , _gm , f) (3), or the probability of observed n _a = a and n _b = b, given the maternal and fetal genotypes, similarly depends on the fetal fraction according to Equation 1. The functional form of F can be a binomial distribution, a beta-binomial distribution, or similar functions as above.
F(a, b, _gc , _gm , f) = P(na = _a , nb = _b | _gc , _gm , f) = P( _na = a, nb = _b | r( _gc , _gm , f)) (3)

In an embodiment, the proportion of offspring can be determined as follows: A maximum likelihood estimate of the fetal fraction f for prenatal testing can be derived without using paternal information. This may be relevant when paternal genetic data is not available, e.g., when the father of record is not in fact the genetic father of the fetus. The fetal fraction is estimated from the set of SNPs where the maternal genotype is 0 or 1, resulting in a set of only two possible fetal genotypes. Define S ₀ as the set of SNPs with maternal genotype 0 and S ₁ as the set of SNPs with maternal genotype 1. The possible fetal genotypes in S ₀ are 0 and 0.5, resulting in a set of possible allele ratios R ₀ (f)={0,f/2}. Similarly, R ₁ (f)={1-f/2,1}. This method can be trivially extended to include SNPs with a maternal genotype of 0.5, but these SNPs are less informative due to the larger set of possible allele ratios.

Define N _a0 and N _b0 as the vectors formed by n _as and n _bs for the SNPs in _S0 , and N _a1 and N _b1 are similarly defined for _S1 . The maximum likelihood estimate of f
is defined by Equation 4.

Assuming that the allele count at each SNP is conditioned independently on the SNP's plasma allele ratio, the probability can be expressed as a product over the SNPs in each set (5).

The dependence on f is due to the set of possible allele ratios R ₀ (f) and R ₁ (f). The SNP probability P(n _as , n _bs |f) can be estimated by assuming the most likely genotype conditioned on f. Selection of the most likely genotype at a reasonably high fetal fraction and read depth will be more reliable. For example, consider a SNP where the mother has genotype zero at a fetal fraction of 10 percent and a read depth of 1000. The expected allele ratios are 0 percent and 5 percent, which are easily distinguishable at a sufficiently high read depth. Substituting the estimated pup genotype into Equation 5 gives the complete Equation (6) for estimating the fetal fraction.

The fetal fraction must be in the range [0,1], and therefore optimization can be easily performed by a constrained one-dimensional search.

In the case of low read depth or in the presence of high noise levels, it may be preferable not to assume the most likely genotype, which may lead to an artificially high confidence. Another method is to sum over the possible genotypes at each SNP, resulting in the following equation (7) for P(n _a ,n _b |f) for the SNPs in S ₀ . The prior probability P(r) can be assumed to be uniform over R ₀ (f) or may be based on population frequencies. Extension to group S ₁ is trivial.

In some embodiments, the probability can be derived as follows: Confidence can be calculated from the data likelihood of the two hypotheses _Ht and _Hf . The likelihood of each hypothesis is derived based on the response model, the estimated fetal fraction, the maternal genotype, the population frequency of the allele, and the plasma allele count.
We define the following notation:
_Gm , _Gc the true maternal genotype and the child's genotype _Gaf , _Gtf the true genotype of the alleged father and the true genotype of the true father G( _gc , _gm , _gtf ) = P( _Gc = _gc | _Gm = _gm , _Gtf = _gtf ) Genetic probability P(g) = P( _Gtf = g) Population frequency of genotype g at a particular SNP

Assuming that the observations at each SNP are independently conditioned on the ratio of plasma alleles, the likelihood of the paternity hypothesis is the product of the likelihoods at the SNPs. The following formula derives the likelihood for a single SNP: Equation 8 is a general expression for the likelihood of any hypothesis h, which is then broken down into the specific cases of _Ht and _Hf .
In the case of H _t , the alleged father is the true father and the fetus's genotype is inherited from the maternal genotype and the alleged father's genotype according to Equation 9.

In the case of H _f , the alleged father is not the true father. The best estimate of the true father's genotype comes from the population frequencies at each SNP. Therefore, the probability of the child's genotype is determined by the known mother's genotype and the population frequencies as in Equation 10.

The confidence in correct paternity, _Cp , is calculated from the SNP-wise product of the two likelihoods using Bayes' rule (11).

Maximum Likelihood Model with Percent Fetal Fraction Determining the ploidy status of a fetus by measuring the free floating DNA contained in maternal serum or by measuring genotypic material in any mixed sample is a non-trivial task. There are several ways, for example, methods that perform read count analysis where the assumption is that if the fetus is trisomic in a particular chromosome, the overall amount of DNA from that chromosome found in the maternal blood will be elevated relative to the reference chromosome. One way to detect trisomy in such fetuses is to normalize the amount of DNA expected for each chromosome, for example, according to the number of SNPs in the analysis set that correspond to a given chromosome, or according to the number of uniquely mappable parts of the chromosome. Once the measurements are normalized, any chromosome with a measured amount of DNA above a certain threshold is determined to be trisomic. This approach is described in Fan et al. PNAS 2008;105(42);16266-16271 and also in Chiu et al. BMJ 2011;342:c7401. In the Chiu et al. paper, normalization was achieved by calculating the Z-score as follows:
Z score for percentage of chromosome 21 in test cases = ((percentage of chromosome 21 in test cases) - (mean percentage of chromosome 21 in reference controls)) / (standard deviation of percentage of chromosome 21 in reference controls).
These methods use a single hypothesis rejection method to determine the ploidy state of the fetus. However, they suffer from several significant drawbacks. These methods for determining ploidy in the fetus are invariant according to the percentage of fetal DNA in the sample, so they use a single cutoff value, resulting in a less than optimal accuracy of the determination, with the worst accuracy occurring when the percentage of fetal DNA in the mixture is relatively low.

In an embodiment, the method of the present disclosure used to determine the ploidy status of a fetus includes taking into account the percentage of fetal DNA in the sample. In another embodiment of the present disclosure, the method includes the use of maximum likelihood estimation. In an embodiment, the method of the present disclosure includes calculating the percent of DNA of fetal or placental origin in the sample. In an embodiment, the threshold for calling aneuploidy is adaptively adjusted based on the calculated percent fetal DNA. In some embodiments, a method for estimating the percentage of DNA that is of fetal origin in a mixture of DNA includes obtaining a mixed sample containing genetic material from the mother and genetic material from the fetus, obtaining a genetic sample from the father of the fetus, measuring the DNA in the mixed sample, measuring the DNA in the father's sample, and calculating the percentage of DNA that is of fetal origin in the mixed sample using the DNA measurements of the mixed sample and the DNA measurements of the father's sample.

In some embodiments of the present disclosure, the proportion of fetal DNA or the percentage of fetal DNA in the mixture can be measured. In some embodiments, the proportion can be calculated using only genotyping measurements made on the maternal plasma sample itself, which is a mixture of fetal and maternal DNA. In some embodiments, the proportion can also be calculated using the measured or otherwise known genotype of the mother and/or the measured or otherwise known genotype of the father. In some embodiments, measurements made on the mixture of maternal and fetal DNA can be used together with knowledge of the parental status to calculate the percent fetal DNA. In some embodiments, the proportion of fetal DNA can be calculated using population frequencies to adjust a model for the probability of certain allele measurements.

In an embodiment of the present disclosure, the confidence in the accuracy of the determination of the ploidy status of the fetus can be calculated. In an embodiment, the confidence of the maximum likelihood (H _major ) hypothesis can be calculated as (1-H _major )/Σ(all H). If all the distributions of the hypotheses are known, the confidence of the hypotheses can be determined. If the genotype information of the parents is known, the distribution of all the hypotheses can be determined. If knowledge of the expected distribution of data for euploid fetuses and the expected distribution of data for aneuploid fetuses is known, the confidence of the ploidy determination can be calculated. If the genotype data of the parents is known, these expected distributions can be calculated. In an embodiment, knowledge of the distribution of the test statistic around the normal hypothesis and the abnormal hypothesis can be used to determine the confidence of the call, as well as to refine the threshold to make a more confident call. This is particularly useful when the amount and/or percentage of fetal DNA in the mixture is low. This helps to avoid situations where a fetus that is actually aneuploid is found to be euploid because a test statistic, such as a Z statistic, does not exceed a threshold based on a threshold optimized for a high percentage of fetal DNA present.

In an embodiment, the methods disclosed herein can be used to determine fetal aneuploidy by determining the number of copies of a maternal target chromosome and a fetal target chromosome in a mixture of maternal and fetal genetic material. The method may involve obtaining a maternal tissue that contains both maternal and fetal genetic material, which in some embodiments may be maternal plasma or tissue isolated from maternal blood. The method may also involve obtaining a mixture of maternal and fetal genetic material from the maternal tissue by processing the maternal tissue. The method may also involve distributing the resulting genetic material into a plurality of reaction samples, e.g., performing high-throughput sequencing on the samples, to randomly provide individual reaction samples that contain target sequences from the target chromosome and individual reaction samples that do not contain target sequences from the target chromosome. The method may involve analyzing target sequences of genetic material present or absent in the individual reaction samples to provide a first number of binary results indicating the presence or absence of a suspected euploid fetal chromosome in the reaction sample, and a second number of binary results indicating the presence or absence of a potentially aneuploid fetal chromosome in the reaction sample. For example, either of the numbers of binary results can be calculated by an informatics technique that counts sequence reads that map to a particular chromosome, a particular region of a chromosome, a particular locus, or a set of loci. The method may include normalizing the number of binary events based on the length of the chromosome, the length of the region of the chromosome, or the number of loci in the population. The method may involve calculating an expected distribution of the number of binary results using the first number for the suspected euploid fetal chromosomes in the reaction sample. The method may involve calculating an expected distribution of the number of binary outcomes for fetal chromosomes suspected to be aneuploid in the reaction sample using the first number and the estimated proportion of fetal DNA found in the mixture, for example by multiplying the expected read number distribution of the number of binary outcomes for fetal chromosomes suspected to be euploid by (1+n/2), where n is the estimated fetal fraction. In some embodiments, the sequence reads can be treated with a probabilistic mapping rather than a binary outcome, which provides greater accuracy but requires more computational power. The fetal fraction can be estimated by multiple methods, some of which are described elsewhere in this disclosure. The method may involve using a maximum likelihood approach to determine whether the second number corresponds to a fetal chromosome that is likely to be aneuploid, euploid or aneuploid. The method may include calling the fetal ploidy state, which is the ploidy state that corresponds to the hypothesis that has the greatest likelihood of being correct given the measured data.

Note that the maximum likelihood model can be used to increase the accuracy of any method of determining the ploidy status of a fetus. Similarly, confidence can be calculated for any method of determining the ploidy status of a fetus. The use of the maximum likelihood model improves the accuracy of any method that uses a single hypothesis rejection method to make ploidy determinations. The maximum likelihood model can be used for any method that can calculate the likelihood distribution for both normal and abnormal cases. The use of the maximum likelihood model implies the ability to calculate confidence for ploidy calls.

Further Considerations of the Method In an embodiment, the methods disclosed herein utilize a quantitative measure of the number of independent observations of each allele of a polymorphic locus, where the step of calculating the ratio of alleles is not included. This differs from methods such as some microarray-based methods that provide information about the ratio of the two alleles of a locus, but do not quantify the number of independent observations of either allele. Some methods known in the art may provide quantitative information about the number of independent observations, but only utilize the ratio of alleles, not the quantitative information, in the calculation that results in the determination of ploidy. To illustrate the importance of retaining information about the number of independent observations, consider a sample locus with two alleles, A and B. In the first experiment, 20 alleles A and 20 alleles B are observed, and in the second experiment, 200 alleles A and 200 alleles B are observed. In both experiments, the ratio (A/(A+B)) is equal to 0.5, but the second experiment conveys more information about the certainty of the frequency of alleles A or B than the first experiment. Rather than utilizing allele ratios, the method uses quantitative data to more accurately model the most likely allele frequencies at each polymorphic locus.

In an embodiment, the method builds a genetic model to aggregate measurements from multiple polymorphic loci to better distinguish between trisomy and disomy and also determine the type of trisomy. Additionally, the method incorporates genetic linkage information to enhance the accuracy of the method. This is in contrast to some methods known in the art that average allele ratios across all of the polymorphic loci on a chromosome. The methods disclosed herein explicitly model the expected allele frequency distribution in disomy, as well as trisomies resulting from nondisjunction during meiosis I, nondisjunction during meiosis II, and nondisjunction during early mitosis of fetal development. To illustrate why this is important, in the absence of crossover, nondisjunction during meiosis I would result in trisomy with two different homologs inherited from one parent, and nondisjunction during meiosis II or early mitosis of fetal development would result in two copies of the same homolog from one parent. Each scenario results in a different predicted allele frequency at each polymorphic locus and also at all physically linked loci (i.e. loci on the same chromosome) considered together. Crossing over, which results in the exchange of genetic material between homologs, adds complexity to the inheritance pattern, but the method adapts to this by using genetic linkage information, i.e., recombination rate information and physical distance between loci. To better distinguish between meiosis I nondisjunction and meiosis II or mitotic nondisjunction, the method incorporates an increased probability of crossing over into the model as the distance from the centromere increases. Meiosis II and mitotic nondisjunction can be distinguished by the fact that mitotic nondisjunction generally results in identical or nearly identical copies of one homolog, while the two homologs present after a meiosis II nondisjunction event are often different due to one or more crossing overs during gamete formation.

In an embodiment, the disclosed method is unable to determine parental haplotypes when disomy is assumed. In an embodiment, in the case of trisomy, the method allows a decision to be made regarding the haplotype of one or both parents by using the fact that the plasma takes two copies from one parent and parental phase information cannot be determined by which of the two copies inherited from the parent in question. In particular, a child can inherit either two identical copies of a parent (concordant trisomy) or both copies of a parent (discordant trisomy). At each SNP, the likelihood of a concordant trisomy and the likelihood of a discordant trisomy can be calculated. In ploidy calling methods that do not use a linkage model that takes into account crossovers, the overall likelihood of trisomy is calculated as a simple weighted average of concordant and discordant trisomy across the entire chromosome. However, only in the presence of crossovers can trisomy change from concordant to discordant on a chromosome (and vice versa) due to biological mechanisms that result in segregation errors and crossovers. The method probabilistically takes into account the likelihood of crossovers, resulting in more accurate ploidy calls than methods that do not take into account the likelihood of crossovers.

In some embodiments, a reference chromosome is used to determine the offspring proportion and the amount or probability distribution of noise levels. In some embodiments, the offspring proportion, noise level, and/or probability distribution are determined using only the genetic information available from the chromosome for which the ploidy state is being determined. The method works without a reference chromosome and without fixing a particular offspring proportion or noise level. This is a significant improvement over and distinct from methods known in the art that require genetic data from a reference chromosome to calibrate offspring proportions and chromosome behavior.

In an embodiment where a reference chromosome is not required to determine the fetal fraction, the hypothesis is determined as follows:
Algorithms using a reference chromosome generally assume that the reference chromosome is disomic, and then (a) fix the most likely child fraction and random noise level N based on this assumption and the data for the reference chromosome:
and then converting as LIK(D|H)=LIK(D|H, cfr ^* , N ^* ), or
(b) Based on this assumption and the data on the reference chromosomes, estimate the distribution of offspring proportions and noise levels. In particular, we do not fix to a single value for cfr and N, but assign probabilities p(cfr,N) to a wide range of possible cfr,N values:
p(cfr,N) ~ LIK(D(ref.chrom)|H11,cfr,N) * priorprob(cfr,N)
where priorprob(cfr,N) is the prior probability of a particular child proportion and noise level, determined by prior knowledge and experimentation, possibly made uniform over the range of cfr,N. Then,
Either method above gives good results.

Note that in some cases, it may not be desirable, possible, or feasible to use a reference chromosome. In such cases, it is possible to derive the best ploidy call for each chromosome separately. In particular:
p(cfr,N|H) can be determined as above for each chromosome separately, assuming hypothesis H rather than simply assuming disomy with respect to the reference chromosome. Using this method, it is possible to keep both noise and offspring fraction parameters fixed, to fix either parameter, or to keep both parameters in probabilistic form for each chromosome and each hypothesis.

Measurements of DNA, especially when the amount of DNA is low or when the DNA is mixed with contaminating DNA, are noisy and/or prone to errors. This noise results in less accurate genotype data and less accurate ploidy calls. In some embodiments, platform modeling or some other noise modeling method can be used to account for the deleterious effects of noise on ploidy determinations. The method uses a joint model of both channels and accounts for random noise due to the amount of input DNA, the quality of the DNA, and/or the quality of the protocol.

This is in contrast to some methods known in the art that use the ratio of allele intensities at loci to make ploidy determinations. This method precludes accurate SNP noise modeling. In particular, errors in measurements are generally not specifically dependent on the measured channel intensity ratios, reducing the model to using one-dimensional information. Accurate modeling of noise, channel quality, and channel interactions requires a simultaneous two-dimensional model that cannot be modeled using allele ratios.

In particular, projecting the two channel information onto a ratio r where f(x,y) is r=x/y does not in itself lend itself to accurate channel noise and bias modeling. The noise at a particular SNP is not a function of the ratio, i.e., noise(x,y)≠f(x,y), but is in fact a joint function of both channels. For example, in a binomial model, the noise of a measured ratio has a variance of r(1-r)/(x+y), which is not purely a function of r. In a model where any channel bias or noise is included, assume that at SNPi, the observed channel X value is x=a _i X+b _i , where X is the true channel value and b _i is the extra channel bias and random noise. Similarly, assume that y=c _i Y+d _i . Because (a _i X+b _i )/(c _i Y+d _i ) is not a function of X/Y, the observed ratio r=x/y cannot accurately predict the true ratio X/Y or model the remaining noise.

The methods disclosed herein describe a method for effective modeling of noise and bias using a joint binomial distribution of all the individual measurement channels. The relevant formulas can be found elsewhere in the document in the sections describing the consistent biases per SNP, P(good), P(ref|bad), and P(mut|bad), which effectively adjust for SNP behavior. In an embodiment, the disclosed methods avoid the implementation limitations that rely solely on allele ratios, and instead use a beta binomial distribution that models behavior based on both channel counts.

In an embodiment, the methods disclosed herein allow for calling the ploidy of a gestating fetus from genetic data found in maternal plasma by using all available measurements. In an embodiment, the methods disclosed herein allow for calling the ploidy of a gestating fetus from genetic data found in maternal plasma by using measurements from only a subset of parental contexts. Some methods known in the art only use genetic data measured when the parental context is from the AA|BB context, i.e., when both parents are homozygous at a given locus but have different alleles. One problem with this method is that the proportion of polymorphic loci from the AA|BB context is small, typically less than 10%. In an embodiment of the methods disclosed herein, the method does not use maternal plasma genetic measurements made at loci with parental contexts of AA|BB. In an embodiment, the method uses plasma measurements only for polymorphic loci with parental contexts of AA|AB, AB|AA, and AB|AB.

Some methods known in the art involve averaging the allele ratios from SNPs in the AA|BB context where both parent genotypes are present and claiming to determine the ploidy call from the average allele ratio at these SNPs. This method is significantly inaccurate due to differential SNP behavior. Note that this method assumes that the genotypes of both parents are known. In contrast, in some embodiments, the method uses a simultaneous channel distribution model that does not assume the presence of either parent and does not assume uniform SNP behavior. In some embodiments, the method takes into account the behavior/weighting of different SNPs. In some embodiments, the method does not require knowledge of the genotype of one or both parents. An example of how the method accomplishes this is as follows:

In some embodiments, the log-likelihood of a hypothesis can be determined for each SNP. For a particular SNPi, given a hypothesis H about fetal ploidy and percent fetal DNAcf, the log-likelihood of the observed data D is:
where m is the possible true mother's genotype, f is the possible true father's genotype, m,f∈{AA,AB,BB}, and c is the possible child's genotype given hypothesis H. In particular, for monosomy, c∈{AA,AB,BB}, for disomy, c∈{AAA,AAB,ABB,BBB}, and for trisomy, c∈{AAA,AAB,ABB,BBB}. Note that while the inclusion of parental genotype data will generally result in more accurate ploidy determination, parental genotype data is not required for the method to work well.

Some methods known in the art involve averaging the allele ratios from SNPs where the mother is homozygous but different alleles are measured in the plasma (AA|AB or AA|BB situations) and claiming to determine the ploidy call from the average allele ratio at these SNPs. This method is intended for cases where the paternal genotype is not available. Note that it is questionable how accurately one can claim that the plasma is heterozygous at a particular SNP without the presence of a homozygous opposite paternal BB: in a small proportion of offspring, the apparent presence of allele B may simply be the presence of noise, and furthermore, the apparent absence of B may be a simple allele dropout in the fetal measurements. Moreover, even if the plasma heterozygosity can actually be determined, this method does not allow for the distinction of paternal trisomies. Specifically, for SNPs where the mother is AA and some B is measured in plasma, if the father is GG, the resulting child's genotype will be AGG, with an average ratio of 33% A (child proportion = 100%). However, if the father is AG, the resulting child's genotype may be AGG for concordant trisomy, giving a ratio of 33% A, or AAG for discordant trisomy, giving an average ratio closer to 66% A. Given that many trisomies are on chromosomes with crossovers, the overall chromosomes can be anywhere between no discordant trisomies and all discordant trisomies, and the ratio can vary anywhere between 33-66%. For normal disomy, the ratio should be around 50%. Without the use of linkage models or accurate error models of the average, this method often misses paternal trisomies. In contrast, the methods disclosed herein assign parental genotype probabilities to each parental genotype candidate based on available genotype information and population frequencies, and do not explicitly require parental genotypes. Furthermore, the methods disclosed herein can detect trisomies in the absence or even presence of parental genotype data, and can compensate by using linkage models to identify crossover points from possible concordant to discordant trisomies.

Several methods known in the art claim to average the ratios of alleles from SNPs where both maternal and paternal genotypes are unknown, and to determine a ploidy call from the average ratios at these SNPs. However, no methods are disclosed that accomplish these objectives. The methods disclosed herein allow accurate ploidy calls to be made in such situations, and implementations are disclosed elsewhere in this document that use joint probability maximum likelihood methods, optionally utilizing models for SNP noise and bias, and linkage models.

Some methods known in the art involve averaging allele ratios and purporting to determine a ploidy call from the average allele ratio at one or a small number of SNPs. However, such methods do not utilize the concept of linkage. The methods disclosed herein do not suffer from these drawbacks.

Using sequence length as a prior to determine DNA origin It has been reported that the distribution of sequence lengths differs between maternal and fetal DNA, with fetal DNA generally being shorter. In certain embodiments of the present disclosure, previous knowledge in the form of empirical data can be used to construct a prior distribution of expected lengths for both maternal DNA (P(X|maternal)) and fetal DNA (P(X|fetal)). Given a new unidentified DNA sequence of length x, it is possible to determine the probability that a given sequence of DNA is either maternal or fetal DNA based on the prior likelihood of x considering either maternal or fetal. In particular, if P(x|maternal)>P(x|fetal), then the DNA sequence can be classified as maternal, P(x|maternal)=P(x|maternal)/[(P(x|maternal)+P(x|fetal)], and if p(x|maternal)<p(x|fetal), then the DNA sequence can be classified as fetal, P(x|fetal)=P(x|fetal)/[(P(x|maternal)+P(x|fetal)]. In an embodiment of the present disclosure, a distribution of maternal and fetal sequence lengths that are specific for the sample can be determined by taking into account sequences that can be assigned with high probability to maternal or fetal, and the distribution specific to the sample can then be used as the expected size distribution for the sample.

Variable read depth to minimize the cost of sequencing Many clinical trials for diagnostics, such as Chiu et al. BMJ, 2011:342:c7401, set up a protocol with several parameters and then run the same protocol with the same parameters for each patient in the trial. When sequencing is used as a method to measure genetic material to determine the ploidy status of a fetus while the mother is pregnant, one relevant parameter is the number of reads. The number of reads can refer to the actual number of reads, the number of intended reads, a split lane of the sequencer, a complete lane, or a complete flow cell. In these trials, the number of reads is generally set at a level that ensures that all or nearly all samples achieve the desired level of accuracy. Sequencing is currently a costly technology, costing approximately $200 per 5 million mappable reads, while as prices come down, any method that allows for sequencing-based diagnostics that operate at a similar level of accuracy but with fewer reads will surely save a significant amount of money.

The accuracy of ploidy determination generally depends on a number of factors, including the number of reads and the percentage of fetal DNA in the mixture. Accuracy is generally higher the higher the percentage of fetal DNA in the mixture. At the same time, accuracy is generally higher the higher the number of reads. It is possible to have a situation with two cases of determining ploidy state with comparable accuracy, where the first case has a smaller percentage of fetal DNA in the mixture than the second case, and where more reads are sequenced in the first case than in the second case. The estimated percentage of fetal DNA in the mixture can be used as a guide in determining the number of reads required to achieve a given level of accuracy.

In certain embodiments of the present disclosure, a collection of samples can be performed where different samples in the collection are sequenced to different read depths, and the number of reads performed for each of the samples is selected to achieve a given level of accuracy, taking into account the calculated fetal DNA percentage in each mixture. In certain embodiments of the present disclosure, this may involve performing a measurement of the mixed sample to determine the percentage of fetal DNA in the mixture, and this estimation of the fetal fraction can be performed using sequencing, can be performed using TAQMAN, can be performed using qPCR, can be performed using SNP arrays, or any method that can distinguish different alleles at a given locus. The need to estimate the fetal fraction can be eliminated by including a hypothesis that encompasses the collection of all or selected fetal fractions in the set of hypotheses considered in the comparison with the actual measurement data. After determining the percentage of fetal DNA in the mixture, the number of sequences to be read for each sample can be determined.

In one embodiment of the present disclosure, 100 pregnant women visit their OB and each person's blood is drawn into a blood tube containing anti-lysant and/or DNAase inactivating. Each woman takes home a kit for the father of the fetus she is carrying to provide a saliva sample. The collection of genetic material from both partners for all 100 couples is sent back to the laboratory where the mother's blood is spun down and the buffy coat is isolated as well as the plasma. The plasma contains a mixture of maternal DNA as well as DNA derived from the placenta. The maternal buffy coat and paternal blood are genotyped using SNP arrays and the DNA in the maternal plasma sample is targeted with SURESELECT hybridization probes. The DNA pulled down with the probes is used to generate 100 tagged libraries, one for each of the maternal samples, with each sample tagged with a different tag. A portion from each library was removed and each of the portions mixed together and added in a multiplexed fashion to two lanes of an ILLUMINA HISEQ DNA sequencer, resulting in approximately 50 million mappable reads from each lane, resulting in approximately 100 million mappable reads in 100 multiplexed mixtures, or approximately 1 million reads per sample. The sequence reads were used to determine the percentage of fetal DNA in each mixture. Fifty samples had more than 15% fetal DNA in the mixture, and 1 million reads were sufficient to determine the fetal ploidy state with 99.9% confidence.

Of the remaining mixtures, 25 had between 10% and 15% fetal DNA, and a portion of each of the relevant libraries prepared from these mixtures was multiplexed and run on one lane of the HISEQ, generating an additional 2 million reads for each sample. The two sets of sequence data for each of the mixtures with between 10% and 15% fetal DNA were added together, and the resulting 3 million reads per sample were sufficient to determine their fetal ploidy state with 99.9% confidence.

Of the remaining mixtures, 13 had between 6% and 10% fetal DNA, and each fraction of the relevant libraries prepared from these mixtures were multiplexed and run on one lane of the HISEQ, generating an additional 4 million reads for each sample. The two sets of sequence data for each of the mixtures with between 6% and 10% fetal DNA were added together, and the resulting 5 million total reads per mixture were sufficient to determine their fetal ploidy state with 99.9% confidence.

Of the remaining mixtures, eight had between 4% and 6% fetal DNA, and a portion of each of the relevant libraries prepared from these mixtures was multiplexed and run on one lane of the HISEQ, generating an additional 6 million reads for each sample. The two sets of sequence data for each of the mixtures with between 4% and 6% fetal DNA were added together, and the resulting 7 million total reads per mixture were sufficient to determine their fetal ploidy state with 99.9% confidence.

Of the remaining four mixtures, all of which had between 2% and 4% fetal DNA, a portion of each of the relevant libraries prepared from these mixtures was multiplexed and run on one lane of the HISEQ, generating an additional 12 million reads for each sample. The two sets of sequence data for each of the mixtures with between 2% and 4% fetal DNA were added together, and the resulting total of 13 million reads per mixture was sufficient to determine their fetal ploidy state with 99.9% confidence.

This method required 6 lanes of sequencing on the HISEQ machine to achieve 99.9% accuracy across 100 samples. If the same number of runs were required for every sample, 25 lanes of sequencing would be taken to ensure every ploidy determination was made with 99.9% accuracy, whereas 14 lanes of sequencing could be achieved if a 4% no-call ratio or error rate was allowed.

Use of raw genotyping data There are a number of ways that NPD can be achieved using fetal genetic information measured in fetal DNA found in maternal blood. Some of these methods involve using SNP arrays to measure fetal DNA, some involve non-targeted sequencing, and some involve targeted sequencing. Targeted sequencing can target SNPs, STRs, other polymorphic loci, non-polymorphic loci, or some combination thereof. Some of these methods may involve using commercial or proprietary allele callers that call the identity of alleles from the intensity data provided by the sensor of the machine that performs the measurement. For example, the ILLUMINA INFINIUM system or the AFFYMETRIX GENECHIP microarray system includes beads or microchips with DNA sequences attached that can hybridize with complementary segments of DNA, and upon hybridization, the fluorescence of the sensor molecule changes, which can be detected. There are also sequencing methods, such as ILLUMINA SOLEXA GENOME SEQUENCER or ABI SOLID GENOME SEQUENCER, which sequence a fragment of DNA for genetic sequence, and when the strand of DNA complementary to the strand to be sequenced is extended, the identity of the extended nucleotide is generally detected via a fluorescent or radioactive tag attached to the complementary nucleotide.In all of these methods, genotype or sequencing data is generally determined based on fluorescence or other signals or their absence.These systems are generally combined with low-level software packages that make specific allele calls (secondary genetic data) from the analog output of fluorescence or other detection devices (primary genetic data).For example, for a given allele on a SNP array, the software makes a call that a particular SNP is present or absent if the fluorescence intensity is above or below a certain threshold amount. Similarly, the output of the sequencer is a chromatogram showing the level of fluorescence detected for each of the dyes, and the software calls that a particular base pair is A or T or C or G. A series of such measurements, commonly referred to as reads, are made by the high-throughput sequencer to indicate the most likely structure of the sequenced DNA sequence. The direct analog output of the chromatogram is defined herein as the primary genetic data, and the base pair/SNP calls made by the software are considered herein as the secondary genetic data. In an embodiment, primary data refers to raw intensity data that is the unprocessed output of a genotyping platform, and the genotyping platform may refer to a SNP array or a sequencing platform. Secondary genetic data refers to processed genetic data where alleles have been called or sequence data has been assigned to base pairs and/or sequence reads have been mapped to the genome.

Many higher level applications leverage these allele calls, SNP calls and sequence reads, i.e. secondary genetic data generated by genotyping software. For example, DNA NEXUS, ELAND or MAQ take sequencing reads and map them to the genome. For example, in the context of non-invasive prenatal diagnosis, complex informatics, such as PARENTAL SUPPORT™, can affect the calling of a large number of SNPs to determine the genotype of an individual. Also, in the context of preimplantation genetic diagnosis, it may be possible to take a set of sequence reads that are mapped to the genome, and by taking normalized read numbers that are mapped to each chromosome or section of a chromosome, it may be possible to determine the ploidy state of an individual. In the context of non-invasive prenatal diagnosis, it may be possible to take a set of sequence reads measured in DNA present in maternal plasma and map them to the genome. Then, it may be possible to take normalized read numbers that are mapped to each chromosome or section of a chromosome, and use that data to determine the ploidy state of an individual. For example, it may be possible to conclude that a chromosome with a disproportionate number of reads is trisomic in the fetus during the pregnancy of the mother whose blood was drawn.

However, in reality, the first output from the measurement instrument is an analog signal. When a particular base pair is called by software associated with the sequencing software, for example software that can call the base pair T, the call is in fact the call that the software considers to be the most likely. However, in some cases, the call may be of low confidence, for example, the analog signal may indicate that a particular base pair is only 90% likely to be T and 10% likely to be A. In another example, the genotype calling software associated with the SNP array reader may call a particular allele to be G. However, in reality, the underlying analog signal may indicate that the allele is only 70% likely to be G and 30% likely to be T. In these cases, some information is lost when using the genotype and sequence calls made by lower level software in higher level applications. That is, the primary genetic data measured directly by the genotyping platform may be messier than the secondary genetic data determined by the accompanying software package, but it contains more information. In mapping the secondary genetic data sequence to the genome, many reads are discarded because some bases are not read clearly enough or the mapping is not clear. When using the primary genetic data sequence reads, all or many of the reads that may have been discarded when they were first converted to secondary genetic data sequence reads can be used by probabilistically processing the reads.

In an embodiment of the present disclosure, the higher level software does not rely on allele calls, SNP calls or sequence reads determined by lower level software. Instead, the higher level software bases its calculations on analog signals measured directly from the genotyping platform. In an embodiment of the present disclosure, an informatics-based method, e.g., PARENTAL SUPPORT™, is modified such that its ability to reconstruct embryo/fetus/child genetic data is engineered to directly use primary genetic data measured by the genotyping platform. In an embodiment of the present disclosure, an informatics-based method, e.g., PARENTAL SUPPORT™, can make allele calls and/or chromosome copy number calls using primary genetic data and without secondary genetic data. In an embodiment of the present disclosure, gene calls, SNP calls, all sequence reads, sequence mapping are probabilistically processed by using raw intensity data measured directly by the genotyping platform rather than converting primary genetic data to secondary gene calls. In one embodiment, the DNA measurements from the prepared sample used in calculating the allele count probabilities and determining the relative probability of each hypothesis comprise primary genetic data.

In some embodiments, the method can increase the accuracy of genetic data of a target individual that incorporates genetic data of at least one related individual, the method comprising obtaining primary genetic data specific to the genome of the target individual and genetic data specific to the genome(s) of the related individual(s), generating a set of one or more hypotheses regarding which segments of chromosomes from the related individual(s) are likely to correspond to those segments in the genome of the target individual, determining the probability of each of the hypotheses given the primary genetic data of the target individual and the genetic data of the related individual(s), and using the probability associated with each hypothesis to determine the most likely state of the actual genetic material of the target individual. In some embodiments, the method allows for determining the number of copies of a segment of a chromosome in the genome of a target individual, the method comprising the steps of: generating a set of copy number hypotheses regarding how many copies of the chromosomal segment are present in the genome of the target individual; incorporating primary genetic data from the target individual and genetic information from one or more related individuals into a dataset; estimating platform response characteristics associated with the dataset, where the platform response may vary from one experiment to another; calculating the relative probability of each of the copy number hypotheses given the dataset and the platform response characteristics; and determining the copy number of the chromosomal segment based on the most likely copy number hypothesis. In an embodiment, a method of the present disclosure can determine the ploidy state of at least one chromosome of a target individual, the method comprising the steps of obtaining primary genetic data from the target individual and from one or more related individuals; generating a set of hypotheses for at least one ploidy state for each of the chromosomes of the target individual; determining, using one or more expert techniques, the statistical probability of each of the hypotheses for the ploidy state in the set; combining, for each of the expert techniques used, the statistical probability determined by the one or more expert techniques for each of the hypotheses for the ploidy state taking into account the obtained genetic data; and determining, for each of the chromosomes of the target individual, the ploidy state based on the combined statistical probability for each of the hypotheses for the ploidy state. In an embodiment, the method of the present disclosure can determine the state of an allele in a set of alleles, in a target individual, and from one or both parents of the target individual, and optionally from one or more related individuals, the method comprising the steps of obtaining primary genetic data from the target individual, from one or both parents, and from any related individuals; generating a set of hypotheses for at least one allele for the target individual, and for one or both parents, and optionally for one or more related individuals, the hypotheses describing possible allele states in the set of alleles; determining a statistical probability for the hypothesis for each allele in the set of hypotheses in light of the obtained genetic data; and determining the allele state for each of the alleles in the set of alleles, for the target individual, and for one or both parents, and optionally for one or more related individuals, based on the statistical probability of each of the hypotheses for the allele.

In some embodiments, the genetic data of the mixed sample may include sequence data, where the sequence data may not be uniquely mapped to the human genome. In some embodiments, the genetic data of the mixed sample may include sequence data, where the sequence data is mapped to multiple locations in the genome, where each possible mapping is accompanied by a probability that the given mapping is correct. In some embodiments, the sequence reads are not assumed to be associated with a specific location in the genome. In some embodiments, the sequence reads are associated with multiple locations in the genome, with an associated probability of belonging to that location.

Counting Methods for Determining Chromosome Copy Number In one aspect, the invention features a method for examining fetal chromosomal abnormality distribution by comparing the number of sequence tags aligned to different chromosomes (see, e.g., U.S. Patent No. 8,296,076, filed April 20, 2012, which is incorporated by reference in its entirety herein). As known in the art, the term "sequence tag" refers to a relatively short (e.g., 15-100) nucleic acid sequence that can be used to identify a given larger sequence that maps to, for example, a chromosome or genomic region or gene. In some embodiments, the method includes the steps of: (i) contacting a sample comprising a mixture of maternal and fetal DNA with a primer library that simultaneously hybridizes to at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different target loci to generate a reaction mixture, wherein the target loci are from a plurality of different chromosomes, the plurality of different chromosomes including at least one first chromosome suspected of having an abnormal distribution in the sample and at least one second chromosome suspected of having a normal distribution in the sample. (ii) subjecting the reaction mixture to primer extension reaction conditions to produce amplification products; (iii) sequencing the amplification products to obtain a plurality of sequence tags of sufficient length to assign specific target loci that align to the target loci; (iv) assigning, by a computer, the plurality of sequence tags to their corresponding target loci; (v) determining, by a computer, the number of sequence tags that align to the target loci of the first chromosome and the number of sequence tags that align to the target loci of the second chromosome; and (vi) comparing the numbers from step (v) to determine the presence or absence of an abnormal distribution of the first chromosome.

In one aspect, the present invention provides a method for detecting the presence or absence of fetal aneuploidy by comparing the relative frequency of target amplification products between chromosomes (see, e.g., International Publication No. WO 2012/103031, filed January 23, 2012, which is incorporated herein by reference in its entirety). In some embodiments, the method includes the steps of: (i) contacting a sample comprising a mixture of maternal and fetal DNA with a library of primers that simultaneously hybridize to at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different non-polymorphic target loci from a plurality of different chromosomes to generate a reaction mixture; (ii) subjecting the reaction mixture to primer extension reaction conditions to generate an amplification product comprising a target amplicon; (iii) quantifying by a computer the relative frequency of the target amplicons from the first and second chromosomes of interest; (iv) comparing by a computer the relative frequency of the target amplicons from the first and second chromosomes of interest; and (v) identifying the presence or absence of an aneuploidy based on the relative frequency of the first and second chromosomes of interest. In some embodiments, the first chromosome is a chromosome suspected to be euploid. In some embodiments, the second chromosome is a chromosome suspected to be aneuploid.

Combination Methods of Prenatal Diagnosis There are many methods that can be used for prenatal diagnosis or prenatal screening of aneuploidy or other genetic defects. Elsewhere in this document, as well as in U.S. Utility Application No. 11/603,406, filed November 28, 2006; U.S. Utility Application No. 12/076,348, filed March 17, 2008, and PCT Application No. PCT/S09/52730, one method is described in which genetic data of related individuals is used to increase the accuracy with which genetic data of a target individual, such as a fetus, is known or presumed. Another method used for prenatal diagnosis involves measuring the levels of certain hormones in maternal blood that are correlated with abnormalities in various genes. An example of this is the so-called triple test, which measures the levels of several (usually two, three, four or five) different hormones in maternal blood. When multiple methods are used to determine the likelihood of a given outcome and no method is conclusive by itself, the information generated by these methods can be combined to produce a more accurate prediction than any of the individual methods. In the triple test, combining information generated from three different hormones can result in a more accurate prediction of genetic abnormalities than can be predicted by individual hormone levels.

Disclosed herein is a method for making more accurate predictions regarding the genetic status of a fetus, specifically the likelihood of a genetic abnormality in the fetus, which includes combining predictions of genetic abnormalities in the fetus made using various methods. A "more accurate" method may refer to a method for diagnosing an abnormality that has a lower false negative rate at a given false positive rate. In preferred embodiments of the present disclosure, one or more of the predictions are made based on known genetic data about the fetus, and the genetic findings are determined using the PARENTAL SUPPORT™ method, i.e., genetic data of individuals related to the fetus are used to determine the genetic data of the fetus with a higher degree of accuracy. In some embodiments, the genetic data may include the ploidy state of the fetus. In some embodiments, the genetic data may refer to a collection of allele calls in the genome of the fetus. In some embodiments, some of the predictions are made using triple testing. In some embodiments, some of the predictions are made using measurements of other hormone levels in the maternal blood. In some embodiments, predictions made by methods that consider diagnosis can be combined with predictions made by methods that consider screening. In some embodiments, the method includes measuring maternal blood levels of alpha-fetoprotein (AFP). In some embodiments, the method includes measuring maternal blood levels of unconjugated estriol (UE ₃ ). In some embodiments, the method includes measuring maternal blood levels of beta-human chorionic gonadotropin (beta-hCG). In some embodiments, the method includes measuring maternal blood levels of invading trophoblast antigen (ITA). In some embodiments, the method includes measuring maternal blood levels of inhibin. In some embodiments, the method includes measuring maternal blood levels of pregnancy associated plasma protein A (PAPP-A). In some embodiments, the method includes measuring maternal blood levels of other hormones or maternal serum markers. In some embodiments, some of the predictions may be made using other methods. In some embodiments, some of the predictions may be made using a fully integrated test, for example, a test that combines an ultrasound and blood test at about 12 weeks of pregnancy and a second blood test at about 16 weeks. In some embodiments, the method includes measuring fetal nuchal edema (NT). In some embodiments, the method includes using the measured levels of the above-mentioned hormones to make a prediction. In some embodiments, the method includes a combination of the above-mentioned methods.

There are many ways to combine predictions, for example hormone measurements can be converted to multiples of the median (MoM) and then converted to likelihood ratios (LR). Similarly, other measurements can be converted to LRs using mixture models of NT distribution. LRs for NT and biochemical markers can be multiplied by age and pregnancy-related risks to derive risks for various conditions such as trisomy 21. Detection rates (DR) and false positive rates (FPR) can be calculated by taking the proportion with risk above a given risk threshold.

In an embodiment, the method for calling the ploidy state includes combining the relative probability of each of the hypotheses about ploidy determined using a joint distribution model and allele count probabilities with the relative probability of each of the hypotheses about ploidy calculated using a statistical technique selected from other methods of determining a risk score for fetal trisomy, including, but not limited to, read count analysis, comparison of heterozygosity rates, statistics available only using parental genetic information, genotype signal probabilities normalized to a particular parental status, statistics calculated using an estimated fetal fraction in the first sample or prepared sample, and combinations thereof.

Another method may involve the situation with four measured hormone levels where the probability distributions around these hormones are known: p( _x1 , _x2 , _x3 , _x4 |e) for euploidy and p( _x1 , _x2 , x3, _x4 |a) for aneuploidy. The probability distributions for the DNA measurements can then be measured, g(y|e) and g(y|a) for euploidy and aneuploidy, respectively. These can be combined as p( _x1 , _x2 , x3, x4|a)g(y|a) and p( _x1 , x2, _x3 , _x4 |e)g(y|e), assuming independence, given the _euploidy / _aneuploidy assumptions, and then multiplied by the priors p(a) and p(e), _respectively , taking into account maternal age. _The highest one can then be selected.

In one embodiment, one can invoke the central limit theorem and assume that the distribution of g(y|a or e) is Gaussian, and measure the mean and standard deviation by examining a large number of samples. In another embodiment, one can assume that the outcomes are not independent, and collect enough samples to estimate the joint distribution p( _x1 , _x2 , _x3 , _x4 |a or e), taking into account the outcomes.

In some embodiments, the ploidy state for which the target individual is determined to be a ploidy state is associated with the hypothesis with the greatest probability. In some cases, one hypothesis has a normalized composite probability of more than 90%. Each hypothesis is associated with one ploidy state or a set of ploidy states, and the ploidy state associated with the hypothesis whose normalized composite probability is greater than 90% or exceeds some other threshold, e.g., 50%, 80%, 95%, 98%, 99% or 99.9%, can be selected as the threshold required for the hypothesis to be called as the determined ploidy state.

DNA from a child from a previous pregnancy in maternal blood One difficulty in non-invasive prenatal testing is to distinguish fetal cells from a current pregnancy from fetal cells from a previous pregnancy. Some believe that genetic matter from a previous pregnancy disappears after some time, but no conclusive evidence has been presented. In an embodiment of the present disclosure, using the PARENTAL SUPPORT™ (PS) method and knowledge of the paternal genome, it is possible to determine fetal DNA of paternal origin (i.e., DNA inherited by the fetus from its father) present in maternal blood. This method can utilize phased parental genetic information. It is possible to phase parental genotypes from non-phased genotype information, using grandparental genetic data (e.g., genetic data measured from the grandfather's sperm), or from genetic data from other born children, or from miscarriage samples. HapMap-based phasing or haplotyping of paternal cells can also be used to phase unphased genetic information. Successful haplotyping has been demonstrated by arresting cells in the mitotic phase where the chromosomes are tightly bundled, and using microfluidics to place separate chromosomes into separate wells. In another embodiment, phased parental haplotype data can be used to detect the presence of two or more homologs from the father, meaning that genetic material from more than one child is present in the blood. By focusing on chromosomes predicted to be euploid in the fetus, the possibility that the fetus suffers from trisomy can be ruled out. It can also be determined if the fetal DNA is not from the current father, in which case other methods, such as triple testing, can be used to predict genetic abnormalities.

There may be other sources of fetal genetic material available by methods other than blood sampling. In the case of fetal genetic material available in maternal blood, there are two main categories: (1) whole fetal cells, e.g., fetal nucleated red blood cells or erythroblasts, and (2) free-floating fetal DNA. In the case of whole fetal cells, there is some evidence that fetal cells can persist in maternal blood for long periods of time, and thus it is possible to isolate cells from pregnant women that contain DNA from the child or fetus from a previous pregnancy. There is also evidence that free-floating fetal DNA is cleared from the system within a few weeks. One challenge is how to determine the identity of the individual whose genetic material is contained in the cells, i.e., how to ensure that the genetic material measured is not from a fetus from a previous pregnancy. In certain embodiments of the present disclosure, knowledge of the maternal genetic material can be used to ensure that the genetic material in question is not maternal genetic material. There are a number of ways to achieve this goal, including informatics-based methods such as PARENTAL SUPPORT™, as described in this document or any of the patents referenced herein.

In an embodiment of the present disclosure, blood drawn from a pregnant mother can be separated into a fraction containing free-floating fetal DNA and a fraction containing nucleated red blood cells. The free-floating DNA can be enriched as necessary, and the genotypic information of the DNA can be measured. From the genotypic information measured from the free-floating DNA, knowledge of the maternal genotype can be used to determine aspects of the fetal genotype. These aspects may refer to the ploidy state and/or identity of the set of alleles. Individual nucleated red blood cells can then be genotyped using methods described elsewhere in this document and in other referenced patents, particularly the methods described in the first section of this document. Knowledge of the maternal genome allows for the determination of whether any given single blood cell is genetically maternal. Also, aspects of the fetal genotype determined as described above allow for the determination of whether a single blood cell is genetically derived from a currently pregnant fetus. Essentially, this aspect of the disclosure allows knowledge of the mother's genes, and possibly genetic information from other related individuals, e.g., the father, to be used together with genetic information measured from free floating DNA found in maternal blood to determine whether isolated nucleated cells found in maternal blood are either (a) genetically maternal, (b) genetically derived from a fetus of the current pregnancy, or (c) genetically derived from a fetus from a previous pregnancy.

Determination of prenatal sex chromosome aneuploidy In methods known in the art, those attempting to determine the sex of a gestating fetus from the mother's blood use the fact that fetal free floating DNA (fffDNA) is present in the mother's plasma. If the Y-specific locus can be detected in the maternal plasma, this means that the gestating fetus is male. However, when using methods known in the art, the absence of detection of the Y-specific locus in the plasma does not necessarily guarantee that the gestating fetus is female, since in some cases the amount of fffDNA is too low to ensure that the Y-specific locus is detected in the case of a male fetus.

Presented herein is a novel method that does not require measuring Y-specific nucleic acids, i.e., DNA that is exclusively from paternally derived loci. The previously disclosed Parental Support method uses crossover frequency data, parental genotype data, and informatics techniques to determine the ploidy state of the fetus during pregnancy. Fetal sex is simply the ploidy state of the fetus at the sex chromosomes. A child that is XX is female, and XY is male. The method described herein also allows the determination of the ploidy state of the fetus. Note that sex determination is effectively synonymous with determining the ploidy of the sex chromosomes; in the case of sex determination, the assumption is often made that the child is euploid, and therefore there are fewer possible hypotheses.

The method disclosed herein involves examining loci common to both the X and Y chromosomes to create a baseline for the amount of fetal DNA present that is expected for a fetus. Regions specific to only the X chromosome can then be examined to determine whether the fetus is female or male. In the case of males, it is expected that less fetal DNA from loci specific to the X chromosome will be found than from loci specific to both X and Y. In contrast, it is expected that female fetuses will have the same amount of DNA from each of these groups. The DNA in question can be measured by any technique capable of quantifying the amount of DNA present in a sample, for example, qPCR, SNP arrays, genotyping arrays, or sequencing. For DNA derived exclusively from one individual, it is expected that the following will be found:
If DNA from the fetus is mixed with DNA from the mother, such that the percentage of fetal DNA in the mixture is F and the percentage of maternal DNA in the mixture is M, such that F+M=100%, we would expect to see the following:
If F and M are known, a predicted ratio can be calculated and the observed data compared to the predicted data. If M and F are unknown, a threshold can be selected based on past data. In either case, the amount of DNA measured at both X and Y specific loci can be used as a baseline, and the test for the sex of the fetus can be based on the amount of DNA observed at the locus specific to the X chromosome only. If the amount is lower than the baseline by an amount approximately equal to ½F or by an amount that brings it below a predefined threshold, the fetus is determined to be male, and if the amount is not approximately equal to the baseline or by an amount that brings it below a predefined threshold, the fetus is determined to be female.

In another embodiment, one can look at only loci that are common to the X and Y chromosomes, often referred to as the Z chromosome. The subset of loci on the Z chromosome is generally always A on the X chromosome and B on the Y chromosome. If the SNP from the Z chromosome is found to have the B genotype, the fetus is called male, and if the SNP from the Z chromosome is found to have only the A genotype, the fetus is called female. In another embodiment, one can look at loci that are found only on the X chromosome. A situation such as AA|B is particularly informative, since the presence of B indicates that the fetus has an X chromosome from the father. A situation such as AB|B is also informative, since it is expected that in female fetuses, only half the B will often be found to be present, compared to male fetuses. In another embodiment, one can look at SNPs on the Z chromosome where both the A and B alleles are present on both the X and Y chromosomes, and it is known which SNPs are from the paternal Y chromosome and which are from the paternal X chromosome.

In one embodiment, it is possible to amplify a single base position that is known to vary between the homologous non-recombining (HNR) regions shared by the Y and X chromosomes. The sequence within this HNR region is nearly identical between the X and Y chromosomes. Within this identical region, there is a single base position that is invariant between the X chromosomes and the Y chromosomes within a population, but differs between the X and Y chromosomes. Each PCR assay can amplify sequences from loci present on both the X and Y chromosomes. Within each of the amplified sequences, there is a single base that can be detected using sequencing or some other method.

In one embodiment, fetal sex can be determined from free floating fetal DNA found in maternal plasma, the method comprising some or all of the following steps: 1) designing PCR (either regular PCR or mini-PCR, plus multiplexing if desired) primers that amplify X/Y variant single base positions within the HNR region; 2) obtaining maternal plasma; 3) PCR amplifying targets from maternal plasma using an HNR X/Y PCR assay; 4) sequencing the amplified products; 5) examining the sequence data for the presence of Y alleles within one or more of the amplified sequences. The presence of one or more indicates a male fetus. The absence of all Y alleles from all amplified products indicates a female fetus.

In an embodiment, targeted sequencing can be used to measure DNA in maternal plasma and/or parental genotypes. In an embodiment, any sequences that are clearly originating from paternally provided DNA can be ignored. For example, in the situation AA|AB, the number of A sequences can be counted and all B sequences can be ignored. To determine the heterozygosity rate for the above algorithm, the number of observed A sequences can be compared to the expected number of total sequences for a given probe. There are many ways that the expected number of sequences can be calculated for each probe per sample. In an embodiment, historical data can be used to determine what fraction of all sequence reads belong to each specific probe, and then this empirical fraction can be used in combination with the total number of sequence reads to estimate the number of sequences in each probe. Another approach can target some known homozygous alleles and then use historical data to correlate the number of reads in each probe with the number of reads in known homozygous alleles. For each sample, the number of reads at the homozygous alleles can then be measured, and this measurement, together with the empirically derived associations, can then be used to estimate the number of sequence reads at each probe.

In some embodiments, it is possible to determine the sex of the fetus by combining predictions made by multiple methods. In some embodiments, the multiple methods are selected from the methods described herein. In some embodiments, at least one of the multiple methods is selected from the methods described herein.

In some embodiments, the methods described herein can be used to determine the ploidy status of a gestational fetus. In some embodiments, the ploidy calling method uses loci specific to the X chromosome or loci common to both the X and Y chromosomes, but does not use any Y-specific loci. In some embodiments, the ploidy calling method uses one or more of the following: loci specific to the X chromosome, loci common to both the X and Y chromosomes, and loci specific to the Y chromosome. In some embodiments, when the ratios of sex chromosomes are similar, e.g., 45,X (Turner syndrome), 46,XX (normal female), and 47,XXX (trisomy X), differentiation can be achieved by comparing the allele distributions according to various hypotheses with the expected allele distributions. In another embodiment, this can be achieved by comparing the relative number of sequence reads for the sex chromosomes with one or more reference chromosomes that are assumed to be euploid. Note also that these methods can be extended to include cases of aneuploidy.

Single-gene disease screening In an embodiment, the method for determining the ploidy status of a fetus can be expanded to allow for simultaneous testing for single-gene disorders. Diagnosis of single-gene diseases affects the same targeting approach used for aneuploidy testing, and requires additional specific targets. In an embodiment, single-gene NPD diagnosis is by linkage analysis. In many cases, direct testing of cfDNA samples is unreliable, since the presence of maternal DNA makes it virtually impossible to determine whether the fetus inherited the mother's mutation. Detecting unique paternally derived alleles is less difficult, but is only fully informative if the disease is dominant and carried by the father, thereby limiting the usefulness of this approach. In an embodiment, the method includes PCR or related amplification techniques.

In some embodiments, the method involves phasing abnormal alleles in the parents that have very tightly linked SNPs around them with information from first-degree relatives. Parental Support can then be run on the targeted sequencing data from these SNPs to determine whether the fetus inherited the normal or abnormal homolog from each parent. As long as the SNPs are well linked, the inheritance of the fetal genotype can be determined with great certainty. In some embodiments, the method includes: (a) adding a set of SNP loci to closely flank a particular set of common diseases to our multiplex pool for testing aneuploidy; (b) reliably phase alleles from these added SNPs with normal and abnormal alleles based on genetic data from various relatives; and (c) reconstructing the fetal haplotype or set of phased SNP alleles in inherited maternal and paternal homologs in the region surrounding the disease locus to determine the fetal genotype. In some embodiments, additional probes that reliably bind to disease-linked loci are added to the set of polymorphic loci used for aneuploidy testing.

Because the sample is a mixture of maternal and fetal DNA, it is difficult to reconstruct the fetal diplotype. In some embodiments, the method incorporates information from close relatives to phase SNPs and disease alleles, and then takes into account the physical distance of SNPs and recombination data from location-specific recombination likelihoods and observed data from maternal plasma genetic measurements to obtain the most likely fetal genotype.

In some embodiments, a number of additional probes per disease-linked locus are included in the set of targeted polymorphic loci; the number of additional probes per disease-linked locus may be between 4 and 10, between 11 and 20, between 21 and 40, between 41 and 60, between 61 and 80, or a combination thereof.

Phasing diploid data from parents can be difficult, and there are many ways this can be done. Some are discussed in this disclosure, others are described in more detail in other disclosures (see, e.g., International Publication Nos. WO2009105531, filed Feb. 9, 2009, and WO2010017214, filed Aug. 4, 2009, which are incorporated herein by reference in their entireties). In one embodiment, parental phase can be inferred and identified by measuring haploid tissue from the parents, e.g., by measuring one or more sperm or eggs. In one embodiment, parental phase can be inferred and identified using measured genotype data from first degree relatives, e.g., parents or siblings of parents. In one embodiment, parental phase can be identified by diluting the DNA in one or more wells until there is expected to be about one or less copy of each haplotype in each well, and then measuring the DNA in one or more wells. In one embodiment, parental genotypes can be phased using a computer program that uses population-based haplotype frequencies to infer the most likely phase. In one embodiment, parental genotypes can be phased when phased haplotype data is known for the other parent in addition to phased unspecified genetic data for one or more of the parent's genetic offspring. In some embodiments, the parent's genetic offspring can be one or more embryos, one or more fetuses, and/or born children. Some of these and other methods of phase-specifying one or both parents are disclosed in further detail, for example, in U.S. Patent Publication Nos. 2011/0033862, filed August 19, 2010; 2011/0178719, filed February 3, 2011; 2007/018446, filed November 22, 2006; and 2008/0243398, filed March 17, 2008. These patents are incorporated herein by reference in their entireties.

Genomic reconstruction of a fetus In one aspect, the invention features a method for determining a haplotype of a fetus. In various embodiments, the method allows for determining which polymorphic loci (e.g., SNPs) were inherited by the fetus and for reconstructing (including recombination events) the homologs present in the fetus (and thereby for interpolating sequences between the polymorphic loci). If desired, substantially the entire genome of the fetus can be reconstructed. If some ambiguity (e.g., crossover intervals) remains in the genome of the fetus, this ambiguity can be minimized by analyzing additional polymorphic loci, if desired. In various embodiments, the polymorphic loci are selected to be at a constant density over a range of one or more chromosomes to reduce all ambiguity to a desired level. The method has important applications for detecting polymorphisms or other mutations of interest in a fetus, since it allows for their detection based on linkage (e.g., the presence of linked polymorphic loci in the fetal genome) rather than direct detection of the polymorphisms or other mutations of interest in the fetal genome. For example, if a parent is a carrier of a mutation associated with cystic fibrosis (CF), a nucleic acid sample containing maternal DNA from the mother of the fetus and fetal DNA from the fetus can be analyzed to determine whether the fetal DNA contains a haplotype that carries the CF mutation. In particular, the polymorphic locus can be analyzed to determine whether the fetal DNA contains a haplotype that carries the CF mutation without the need to detect the CF mutation itself in the fetal DNA.

In some embodiments, the method includes determining parental haplotypes (e.g., maternal and paternal haplotypes of the fetus). In some embodiments, this determination is made without using data from maternal or paternal relatives. In some embodiments, the parental haplotypes are determined using a dilution technique followed by SNP genotyping or sequencing, as described herein and elsewhere (see, e.g., U.S. Patent Publication No. 2011/0033862, filed Aug. 19, 2010, which is incorporated herein by reference in its entirety). Because the DNA is diluted, there will not be more than one haplotype in the same portion (or tube). This effectively results in a single molecule of DNA in the tube, and the haplotypes on the single DNA molecule can be determined. In some embodiments, the method includes dividing the DNA sample into multiple portions, with at least one portion comprising a chromosome or a chromosome segment from a chromosome pair, and genotyping the DNA sample in at least one portion (e.g., determining the presence of two or more polymorphic loci), thereby determining parental haplotypes. In some embodiments, the genotyping includes sequencing (e.g., shotgun sequencing). In some embodiments, the genotyping includes using a SNP array to detect polymorphic loci, e.g., at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different polymorphic loci. In some embodiments, the genotyping includes using multiplex PCR. In some embodiments, the method includes contacting the aliquot sample with a library of primers that simultaneously hybridize to at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different polymorphic loci (e.g., SNPs) to generate a reaction mixture, and subjecting the reaction mixture to primer extension reaction conditions to obtain amplification products that are measured on a high-throughput sequencer to generate sequencing data.

In some embodiments, the maternal haplotype is determined by any of the methods described herein using data from the maternal relatives. In some embodiments, the paternal haplotype is determined by any of the methods described herein using data from the paternal relatives. In some embodiments, haplotypes for both the father and the mother are determined. In some embodiments, a SNP array is used to determine the presence of at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 distinct polymorphic loci in DNA samples from the mother (or father) and the maternal (or paternal) relatives. In some embodiments, the method includes contacting a DNA sample from the mother (or father) and/or maternal (or paternal) relatives with a primer library that simultaneously hybridizes to at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different polymorphic loci (e.g., SNPs) to generate a reaction mixture, and subjecting the reaction mixture to primer extension reaction conditions to obtain amplification products that are measured on a high-throughput sequencer to generate sequencing data. Parental haplotypes can be determined based on SNP arrays or sequencing data. In some embodiments, parental data can be phased by methods described or referenced elsewhere in this document.

This parental haplotype data can be used to determine whether the fetus has inherited the parental haplotypes. In some embodiments, a nucleic acid sample containing maternal DNA from the mother of the fetus and fetal DNA from the fetus is analyzed using a SNP array to detect at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 distinct polymorphic loci. In some embodiments, a nucleic acid sample containing maternal DNA from the mother of the fetus and fetal DNA from the fetus is analyzed by contacting the sample with a primer library that simultaneously hybridizes to at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different polymorphic loci (e.g., SNPs) to generate a reaction mixture. In some embodiments, the reaction mixture is subjected to primer extension reaction conditions to generate an amplification product. In some embodiments, the amplification product is measured in a high-throughput sequencer to generate sequencing data. In various embodiments, the SNP array or sequencing data is used to determine parental haplotypes by modeling the dependency between polymorphic alleles on a chromosome using data on chromosomal crossover probability at various sites in the chromosome (e.g., by using recombination data such as can be found in the HapMap database to generate a recombination risk score for any interval). In some embodiments, the allele counts at the polymorphic loci are computed based on the sequencing data. In some embodiments, multiple ploidy hypotheses for different possible chromosomal ploidy states are computed, a model (e.g., a joint distribution model) for the expected allele counts at the polymorphic loci on the chromosome is computed for each ploidy hypothesis, the relative probability of each ploidy hypothesis is computed using the joint distribution model and the allele counts, and the ploidy state of the fetus is called by selecting the ploidy state corresponding to the hypothesis with the greatest probability. In some embodiments, the steps of constructing a joint distribution model for the allele counts and determining the relative probability of each hypothesis are performed using a method that does not require the use of a reference chromosome.

In some embodiments, fetal haplotypes are determined for one or more chromosomes selected from the group consisting of chromosomes 13, 18, 21, X, and Y. In some embodiments, fetal haplotypes are determined for all fetal chromosomes. In various embodiments, the method determines substantially the entire fetal genome. In some embodiments, haplotypes are determined for at least 30, 40, 50, 60, 70, 80, 90, or 95% of the fetal genome. In some embodiments, fetal haplotype determination includes information about which alleles are present for at least 1,000; 2,000; 5,000; 7,500; 10,000; 20,000; 25,000; 30,000; 40,000; 50,000; 75,000; or 100,000 different polymorphic loci.

Composition of DNA When performing informatics analysis on sequencing data measured on a mixture of fetal and maternal blood to determine genomic information about the fetus, for example the ploidy state of the fetus, it may be advantageous to measure the allele distribution in the set of alleles. Unfortunately, in many cases, such as when attempting to determine the ploidy state of the fetus from the DNA mixture found in the plasma of a maternal blood sample, the amount of DNA available is not sufficient to directly measure the allele distribution in the mixture with good fidelity. In these cases, amplifying the DNA mixture provides a sufficient number of DNA molecules that can measure the desired allele distribution with good fidelity. However, current methods of amplification commonly used to amplify DNA for sequencing are often highly biased, i.e., both alleles of a polymorphic locus are not amplified in the same amount. Biased amplification can result in an allele distribution that is significantly different from the allele distribution in the original mixture. For most purposes, it is not necessary to measure the relative amounts of alleles present at a polymorphic locus very accurately. In contrast, in certain embodiments of the present disclosure, amplification or enrichment methods that specifically enrich for polymorphic alleles and preserve the ratio of alleles are advantageous.

Described herein are a number of methods that can be used to preferentially enrich a sample of DNA at multiple loci with minimal allelic bias. In some examples, circularization probes are used to target multiple loci, and the 3' and 5' ends of the pre-circularization probe are designed to hybridize to bases one or a few positions away from the polymorphic site of the targeted allele. Another example is to use PCR probes where the 3' end PCR probe is designed to hybridize to bases one or a few positions away from the polymorphic site of the targeted allele. Another example is to use a split-and-pool approach to create a mixture of DNA where the preferentially enriched loci are enriched with less allelic bias and without the drawbacks of direct multiplexing. Another example is to use a hybrid capture approach where the capture probe is designed such that the region of the capture probe designed to hybridize to DNA adjacent to the targeted polymorphic site is separated from the polymorphic site by one or a few bases.

When using the measured allele distribution at a set of polymorphic loci to determine the ploidy state of an individual, it is desirable to preserve the relative abundance of alleles in a DNA sample when the sample is prepared for genetic measurement. This preparation may involve WGA amplification, targeted amplification, selective enrichment techniques, hybrid capture methods, circularization probes, or other methods intended to amplify the amount of DNA and/or selectively enhance the presence of DNA molecules corresponding to specific alleles.

In some embodiments of the present disclosure, there is a set of DNA probes designed to target loci with the highest minor allele frequency. In some embodiments of the present disclosure, there is a set of probes designed to target loci with the highest likelihood that the fetus has an informative SNP at those loci. In some embodiments of the present disclosure, there is a set of probes designed to target loci optimized for a given population subgroup. In some embodiments of the present disclosure, there is a set of probes designed to target loci optimized for a given mixture of population subgroups. In some embodiments of the present disclosure, there is a set of probes designed to target loci optimized for a given pair of parents from different population subgroups with different minor allele frequency profiles. In some embodiments of the present disclosure, there is a circularized DNA strand that includes at least one base pair annealed to a portion of DNA of fetal origin. In some embodiments of the present disclosure, there is a circularized DNA strand that includes at least one base pair annealed to a portion of DNA of placental origin. In some embodiments of the present disclosure, there is a circularized DNA strand that has been circularized while at least some of the nucleotides are annealed to DNA of fetal origin. In some embodiments of the present disclosure, there is a circularized DNA strand that has been circularized while at least some of the nucleotides are annealed to DNA of placental origin. In some embodiments of the present disclosure, there is a collection of probes, some of which target single tandem repeats and some of which target single nucleotide polymorphisms. In some embodiments, a locus is selected for non-invasive prenatal diagnosis. In some embodiments, a probe is used for non-invasive prenatal diagnosis. In some embodiments, a locus is targeted using a method that may include a circularization probe, a MIP, capture with a hybridization probe, a probe on a SNP array, or a combination thereof. In some embodiments, a probe is used as a circularization probe, a MIP, capture with a hybridization probe, a probe on a SNP array, or a combination thereof. In some embodiments, a locus is sequenced for non-invasive prenatal diagnosis.

In some embodiments, the number of sequence reads containing SNPs with known parental context can be increased by preferentially amplifying specific sequences using qPCR, where the relative informativeness of the sequence is greater when combined with the relevant parental context. In some embodiments, the number of sequence reads containing SNPs with known parental context can be increased by preferentially amplifying specific sequences using circularization probes (e.g., MIPs). In some embodiments, the number of sequence reads containing SNPs with known parental context can be increased by preferentially amplifying specific sequences using capture by hybridization (e.g., SURESELECT). Different methods can be used to increase the number of sequence reads containing SNPs with known parental context. In some embodiments, targeting can be achieved by extension ligation, ligation without extension, capture by hybridization, or PCR.

In a sample of fragmented genomic DNA, a fraction of the DNA sequence maps uniquely to an individual chromosome, while other DNA sequences are found on different chromosomes. It should be noted that DNA found in plasma, whether of maternal or fetal origin, is generally fragmented to lengths often less than 500 bp. In a typical genomic sample, approximately 3.3% of the mappable sequences map to chromosome 13, 2.2% of the mappable sequences map to chromosome 18, 1.35% of the mappable sequences map to chromosome 21, 4.5% of the mappable sequences map to the X chromosome in females, 2.25% of the mappable sequences map to the X chromosome in males, and 0.73% of the mappable sequences map to the Y chromosome in males. These are the chromosomes most likely to be aneuploid in the fetus. Also, using the SNPs contained in dbSNP, approximately 1 in 20 sequences in short sequences contain a SNP. This percentage could be even higher, given that there may be many SNPs that remain to be discovered.

In certain embodiments of the present disclosure, the targeting method can be used to enhance the portion of DNA in a sample of DNA that maps to a given chromosome such that the portion significantly exceeds the percentages typical of genomic samples listed above. In certain embodiments of the present disclosure, the targeting method can be used to enhance the portion of DNA in a sample of DNA such that the percentage of sequences that contain SNPs significantly exceeds the percentages that might typically be found in a genomic sample. In certain embodiments of the present disclosure, the targeting method can be used to target DNA from chromosomes or sets of SNPs in a mixture of maternal and fetal DNA for prenatal diagnosis.

It should be noted that methods have been reported for determining fetal aneuploidy by counting the number of reads that map to a suspect chromosome and comparing it to the number of reads that map to a reference chromosome, with the assumption that an excess of reads on a suspect chromosome corresponds to fetal triploidy at that chromosome (U.S. Pat. No. 7,888,017). These methods for prenatal diagnosis do not use any kind of targeting, and the use of targeting for prenatal diagnosis is not described.

By using targeted approaches in sequencing mixed samples, it may be possible to achieve a certain level of accuracy using fewer sequence reads. Accuracy may refer to sensitivity, accuracy may refer to specificity, or accuracy may refer to some combination thereof. The desired level of accuracy may be between 90% and 95%, the desired level of accuracy may be between 95% and 98%, the desired level of accuracy may be between 98% and 99%, the desired level of accuracy may be between 99% and 99.5%, the desired level of accuracy may be between 99.5% and 99.9%, the desired level of accuracy may be between 99.9% and 99.99%, the desired level of accuracy may be between 99.99% and 99.999%, and the desired level of accuracy may be between 99.999% and 100%. A level of accuracy greater than 95% may be referred to as high accuracy.

There are several published prior art methods that demonstrate how fetal ploidy status can be determined from a mixed sample of maternal and fetal DNA, e.g., G. J. W. Liao et al. Clinical Chemistry 2011; 57(1) 92-101. These methods focus on thousands of locations along each chromosome. While the number of locations along chromosomes that can be targeted and, from a mixed sample of NA, yields high accuracy in determining ploidy in the fetus for a given number of sequence reads, is unexpectedly low. In an embodiment of the present disclosure, accurate ploidy determination can be performed by using any targeted method, such as qPCR, ligand-mediated PCR, other PCR methods, hybridization capture or targeted sequencing using circularization probes, where the number of loci along a chromosome that need to be targeted can be between 5,000 and 2,000 loci, between 2,000 and 1,000 loci, between 1,000 and 500 loci, between 500 and 300 loci, between 300 and 200 loci, between 200 and 150 loci, between 150 and 100 loci, between 100 and 50 loci, between 50 and 20 loci, between 20 and 10 loci. Optimally, the number of loci along a chromosome that need to be targeted can be between 100 and 500 loci. A high level of accuracy can be achieved by targeting a small number of loci and performing an unexpectedly small number of sequence reads. The number of reads may be between 100 million and 50 million reads, the number of reads may be between 50 million and 20 million reads, the number of reads may be between 20 million and 10 million reads, the number of reads may be between 10 million and 5 million reads, the number of reads may be between 5 million and 2 million reads, the number of reads may be between 2 million and 1 million; the number of reads may be between 1 million and 500,000; the number of reads may be between 500,000 and 200,000, the number of reads may be between 200,000 and 100,000, the number of reads may be between 100,000 and 50,000, the number of reads may be between 50,000 and 20,000, the number of reads may be between 20,000 and 10,000, the number of reads may be less than 10,000. Fewer reads are needed for larger amounts of input DNA.

In some embodiments, there are compositions comprising a mixture of DNA of fetal and maternal origin, where the percentage of sequences uniquely mapping to chromosome 13 is greater than 4%, greater than 5%, greater than 6%, greater than 7%, greater than 8%, greater than 9%, greater than 10%, greater than 12%, greater than 15%, greater than 20%, greater than 25%, or greater than 30%. In some embodiments of the present disclosure, there are compositions comprising a mixture of DNA of fetal and maternal origin, where the percentage of sequences uniquely mapping to chromosome 18 is greater than 3%, greater than 4%, greater than 5%, greater than 6%, greater than 7%, greater than 8%, greater than 9%, greater than 10%, greater than 12%, greater than 15%, greater than 20%, greater than 25%, or greater than 30%. In some embodiments of the present disclosure, there are compositions comprising a mixture of DNA of fetal and maternal origin, where the percentage of sequences uniquely mapping to chromosome 21 is greater than 2%, greater than 3%, greater than 4%, greater than 5%, greater than 6%, greater than 7%, greater than 8%, greater than 9%, greater than 10%, greater than 12%, greater than 15%, greater than 20%, greater than 25%, or greater than 30%. In some embodiments of the present disclosure, there are compositions comprising a mixture of DNA of fetal and maternal origin, where the percentage of sequences uniquely mapping to chromosome X is greater than 6%, greater than 7%, greater than 8%, greater than 9%, greater than 10%, greater than 12%, greater than 15%, greater than 20%, greater than 25%, or greater than 30%. In some embodiments of the present disclosure, there are compositions that include a mixture of DNA of fetal and maternal origin, where the percentage of sequences that uniquely map to the Y chromosome is greater than 1%, greater than 2%, greater than 3%, greater than 4%, greater than 5%, greater than 6%, greater than 7%, greater than 8%, greater than 9%, greater than 10%, greater than 12%, greater than 15%, greater than 20%, greater than 25%, or greater than 30%.

In some embodiments, the composition is described as comprising a mixture of DNA of fetal and maternal origin, wherein the percentage of sequences uniquely mapping to a chromosome and containing at least one single nucleotide polymorphism is greater than 0.2%, greater than 0.3%, greater than 0.4%, greater than 0.5%, greater than 0.6%, greater than 0.7%, greater than 0.8%, greater than 0.9%, greater than 1%, greater than 1.2%, greater than 1.4%, greater than 1.6%, greater than 1.8%, greater than 2%, greater than 2.5%, greater than 3%, greater than 4%, greater than 5%, greater than 6%, greater than 7%, greater than 8%, greater than 9%, greater than 10%, greater than 12%, greater than 15%, or greater than 20%, and the chromosome is selected from the group of 13, 18, 21, X, or Y. In some embodiments of the present disclosure, a composition comprises a mixture of DNA of fetal and maternal origin, wherein the percentage of sequences that uniquely map to a chromosome and contain at least one single nucleotide polymorphism from a set of single nucleotide polymorphisms is greater than 0.15%, greater than 0.2%, greater than 0.3%, greater than 0.4%, greater than 0.5%, greater than 0.6%, greater than 0.7%, greater than 0.8%, greater than 0.9%, greater than 1%, greater than 1.2%, greater than 1.4%, greater than 1.6%, greater than 1.8%, greater than 2%, greater than 2.5%, greater than 3%, greater than 4%, greater than 5%, greater than 6%, greater than 7%, greater than 8%, greater than 9%, greater than 10%, greater than 12%, greater than 15%, or greater than 20%, and the chromosomes are chromosomes 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 8 There are compositions selected from the set of chromosomes 8, 21, X and Y, in which the number of single nucleotide polymorphisms in the set of single nucleotide polymorphisms is between 1 and 10, between 10 and 20, between 20 and 50, between 50 and 100, between 100 and 200, between 200 and 500, between 500 and 1,000, between 1,000 and 2,000, between 2,000 and 5,000, between 5,000 and 10,000, between 10,000 and 20,000, between 20,000 and 50,000, and between 50,000 and 100,000.

In theory, each cycle of amplification doubles the amount of DNA present, but in practice, the degree of amplification is slightly less than two-fold. In theory, amplification, including targeted amplification, results in the amplification of an unbiased DNA mixture, but in practice, different alleles tend to be amplified to different degrees than other alleles. When amplifying DNA, the degree of allelic bias generally increases with the number of amplification steps. In some embodiments, the methods described herein include amplifying DNA with a low level of allelic bias. Since the allelic bias is compounded with each additional cycle, the allelic bias per cycle can be determined by calculating the n-th root of the overall bias, where n is the logarithm to the base 2 of the degree of enrichment. In some embodiments there is a composition comprising a second mixture of DNA, wherein the second mixture of DNA is preferentially enriched at a plurality of polymorphic loci from the first mixture of DNA, where the degree of enrichment is at least 10, at least 100, at least 1,000, at least 10,000, at least 100,000, or at least 1,000,000, and wherein the ratio of alleles at each locus in the second mixture of DNA differs, on average, from the ratio of alleles at that locus in the first mixture of DNA by a factor of less than 1,000%, 500%, 200%, 100%, 50%, 20%, 10%, 5%, 2%, 1%, 0.5%, 0.2%, 0.1%, 0.05%, 0.02%, or 0.01%. In some embodiments, there is a composition comprising a second mixture of DNA, the second mixture of DNA being preferentially enriched for a plurality of polymorphic loci from the first mixture of DNA, and wherein the allelic bias for the plurality of polymorphic loci per cycle is, on average, less than 10%, 5%, 2%, 1%, 0.5%, 0.2%, 0.1%, 0.05%, or 0.02%. In some embodiments, the plurality of polymorphic loci comprises at least 10 loci, at least 20 loci, at least 50 loci, at least 100 loci, at least 200 loci, at least 500 loci, at least 1,000 loci, at least 2,000 loci, at least 5,000 loci, at least 10,000 loci, at least 20,000 loci, or at least 50,000 loci.

Some embodiments In some embodiments, disclosed herein is a method for generating a report disclosing determined ploidy states for chromosomes in a gestating fetus, the method comprising the steps of obtaining a first sample containing DNA from the mother of the fetus and DNA from the fetus, obtaining genotype data from one or both parents of the fetus, preparing the first sample by isolating the DNA to obtain a prepared sample, measuring the DNA in the prepared sample at a plurality of polymorphic loci, computing an allele count or allele count probability at a plurality of polymorphic loci from the DNA measurements obtained for the prepared sample, and performing staining for the different possible ploidy states of the chromosomes. The method includes generating, on a computer, a plurality of ploidy hypotheses regarding expected allele count probabilities at a plurality of polymorphic loci on the body; for each ploidy hypothesis, constructing, on a computer, a joint distribution model for the allele count probabilities at each polymorphic locus on the chromosome using genotype data from one or both parents of the fetus; determining, on a computer, the relative probability of each of the ploidy hypotheses using the joint distribution model and the calculated allele count probabilities for the prepared samples; calling the ploidy state of the fetus by selecting the ploidy state corresponding to the hypothesis with the greatest probability; and generating a report disclosing the determined ploidy states.

In some embodiments, the method is used to determine the ploidy state of multiple gestating fetuses in multiple respective mothers, the method further comprising determining the percent DNA of fetal origin in each of the prepared samples, where measuring the DNA in the prepared samples is performed by sequencing a number of DNA molecules in each prepared sample, sequencing more DNA molecules from prepared samples having smaller fractions of fetal DNA than prepared samples having larger fractions of fetal DNA.

In some embodiments, the method is used to determine the ploidy state of a plurality of gestating fetuses in a plurality of respective mothers, wherein the step of measuring DNA in the prepared sample is performed by sequencing, for each fetus, a first fraction of the prepared sample of DNA to obtain a first set of measurements, and the method further comprises the steps of making a first relative probability determination for each of the hypotheses about the ploidy of each fetus given the first set of DNA measurements, and determining whether the first relative probability determination for each of the hypotheses about the ploidy is consistent with a ploidy corresponding to an aneuploid fetus. The method further includes resequencing a second fraction of the prepared sample from the fetus that indicates that the hypothesis about the ploidy has a significant but inconclusive probability to obtain a second set of measurements, making a second relative probability determination for the hypothesis about the ploidy of the fetus using the second set of measurements and, optionally, the first set of measurements, and calling the ploidy state of the fetus after resequencing the second sample by selecting the ploidy state corresponding to the hypothesis having the greatest probability as determined by the second relative probability determination.

In some embodiments, a composition is disclosed that includes a preferentially enriched DNA sample, the preferentially enriched DNA sample being preferentially enriched at a plurality of polymorphic loci from a first DNA sample, the first DNA sample being comprised of a mixture of maternal and fetal DNA derived from maternal plasma, the degree of enrichment being at least two-fold, and the allelic bias between the first sample and the preferentially enriched sample being, on average, selected from the group consisting of less than 2%, less than 1%, less than 0.5%, less than 0.2%, less than 0.1%, less than 0.05%, less than 0.02%, and less than 0.01%. In some embodiments, a method for making such preferentially enriched DNA samples is disclosed.

In some embodiments, a method for determining the presence or absence of fetal aneuploidy in a maternal tissue sample comprising fetal genomic DNA and maternal genomic DNA, comprising: (a) obtaining a mixture of fetal and maternal genomic DNA from said maternal tissue sample; (b) selectively enriching the mixture of fetal and maternal DNA at a plurality of polymorphic alleles; (c) distributing the fragments selectively enriched from the mixture of fetal and maternal genomic DNA in step a to result in a reaction sample comprising single genomic DNA molecules or amplification products of single genomic DNA molecules; (d) performing massively parallel DNA sequencing on the fragments of genomic DNA selectively enriched in the reaction sample in step c) to determine the sequences of said selectively enriched fragments; (e) identifying the chromosomes to which the sequences obtained in step d) belong; and (f) analyzing the data from step d) to identify i) chromosomes belonging to at least one initial target chromosome suspected to be diploid in both the mother and the fetus, and The method includes the steps of: (g) calculating a predicted distribution of the number of fragments of genomic DNA from step d) using the number determined in step f) part i) for the second target chromosome if the second target chromosome is euploid; (h) calculating a predicted distribution of the number of fragments of genomic DNA from step d) using the first number, step f) part i), and the estimated proportion of fetal DNA found in the mixture of step b) for the second target chromosome if the second target chromosome is aneuploid; and (i) determining using a maximum likelihood or maximum a posteriori method whether the number of fragments of genomic DNA determined in step f) part ii) is more likely to be part of the distribution calculated in step g) or the distribution calculated in step h), thereby indicating the presence or absence of fetal aneuploidy.

Representative Cancer Diagnostic Methods It is noted that it has been demonstrated that DNA originating from a cancer surviving in a host can be found in the blood of the host. Just as genetic diagnosis can be performed by measuring the mixed DNA found in maternal blood, genetic diagnosis can equally well be performed by measuring the mixed DNA found in host blood. Genetic diagnosis can include aneuploidy status or genetic mutation. Any claims in the present disclosure that read on determining the ploidy status or genetic status of a fetus from measurements made on maternal blood can equally well be read on determining the ploidy status or genetic status of a cancer from measurements made on host blood.

In some embodiments, the disclosed method allows for the determination of the ploidy state of a cancer, the method comprising obtaining a mixed sample containing genetic material from a host and genetic material from a cancer, measuring DNA in the mixed sample, calculating the percentage of DNA of cancer origin in the mixed sample, and determining the ploidy state of the cancer using the measurements obtained for the mixed sample and the calculated percentage. In some embodiments, the method may further comprise administering a cancer treatment based on the determination of the ploidy state of the cancer. In some embodiments, the method may further comprise administering a cancer treatment based on the determination of the ploidy state of the cancer, the cancer treatment being selected from the group including pharmaceuticals, biological therapeutics, and antibody-based therapies and combinations thereof.

Exemplary Implementation Methods Any of the embodiments disclosed herein may be implemented in digital electronic circuitry, integrated circuits, specially designed ASICs (Application Specific Integrated Circuits), computer hardware, firmware, software, or in combinations thereof. The apparatus of the presently disclosed embodiments may be implemented in a computer program product tangibly embodied in a machine-readable storage device for execution by a programmable processor, and the steps of the methods of the presently disclosed embodiments may be performed by the programmable processor executing a program of instructions for performing the functions of the presently disclosed embodiments by manipulating input data and generating output. The presently disclosed embodiments may be advantageously implemented in one or more computer programs executable and/or interpretable in a programmable system including at least one programmable processor in conjunction with a storage system for receiving and transmitting data and instructions, at least one input device, and at least one output device, which may be special or general purpose. Each computer program may be implemented in a high-level procedural or object-oriented programming language, or in assembly or machine language, if desired, and in any case the language may be a compiled or interpreted language. A computer program may be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment. A computer program may be deployed to be executed or interpreted on one computer or on multiple computers at a single site or distributed across multiple sites and interconnected by a communications network.

Computer-readable storage media, as used herein, refers to physical or tangible storage devices (as opposed to signals), including, without limitation, volatile and non-volatile, removable and non-removable media implemented in any method or technology for the tangible storage of information, such as computer-readable instructions, data structures, program modules or other data. Computer-readable storage media include, but are not limited to, RAM, ROM, EPROM, EEPROM, flash memory or other solid-state storage technology, CD-ROM, DVD or other optical storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices or any other physical or tangible media that can be used to tangibly store desired information or data or instructions and that can be accessed by a computer or processor.

Any of the methods described herein may include outputting data in a physical format, such as a computer screen or printout on paper. In the description of any of the embodiments elsewhere in this document, it should be understood that the described methods may be combined with outputting ready-to-use data in a format on which a physician can act. Additionally, the described methods may be combined with the actual performance of a clinical decision resulting in the performance of a clinical treatment, or a clinical decision not to act. Some of the embodiments described herein for determining genetic data related to a target individual may be combined in the context of IVF with the decision to select one or more embryos for transfer, and optionally with the process of transferring the embryos into the uterus of a future mother. Some of the embodiments described herein for determining genetic data related to a target individual may be combined with notification by a medical professional of a potential chromosomal abnormality or absence thereof, and optionally with the decision to abort or not abort the fetus, in the context of prenatal diagnosis. Some of the embodiments described herein may be combined with outputting ready-to-use data and performing a clinical decision resulting in the performance of a clinical treatment, or a clinical decision not to act.

Exemplary Diagnostic Box In an embodiment, the present disclosure includes a diagnostic box capable of partially or completely performing any of the methods described herein. In an embodiment, the diagnostic box may be located in a doctor's office, a hospital laboratory, or any suitable location reasonably close to the patient care location. The box may allow the entire method to be performed in a fully automated manner, or the box may require one or several steps to be completed manually by a technician. In an embodiment, the box may allow the genotype data measured at least for the maternal plasma to be analyzed. In an embodiment, the box may be coupled with a means of transmitting the genotype data measured in the diagnostic box to an external computing facility that then analyzes the genotype data and possibly also generates a report. The diagnostic box may include a robotic unit capable of transferring aqueous or liquid samples from one container to another. The diagnostic box may include any number of reagents, both solid and liquid. The diagnostic box may include a high throughput sequencer. The diagnostic box may include a computer.

Experimental Section The embodiments disclosed herein are described in the following examples, which are provided to aid in the understanding of the present disclosure and should not be construed in any way as limiting the scope of the present disclosure as defined in the claims that follow. The following examples are presented to provide those skilled in the art with a complete disclosure and description of how to use the described embodiments, and are not intended to limit the scope of the disclosure, nor are they intended to represent that the following experiments are all or the only experiments performed. Although attempts have been made to ensure accuracy with respect to numbers used (e.g., amounts, temperatures, etc.), some experimental error and deviation should be accounted for. Unless otherwise specified, parts are parts by volume and temperatures are in degrees Celsius. It should be understood that variations in the methods described can be made without changing the fundamental aspects that the experiments are meant to illustrate.

Experiment 1
The objective was to show that a Bayesian maximum likelihood estimation (MLE) algorithm that uses parental genotypes to calculate the fetal fraction improves the accuracy of non-invasive prenatal trisomy diagnosis compared with published methods.
Simulated sequencing data for maternal cfDNA was generated by sampling the reads obtained in trisomy 21 and each maternal cell line. The rate of correct disomy and trisomy calls was determined from 500 simulations in various fetal fractions for published methods (Chiu et al. BMJ 2011:342:c7401) and our MLE-based algorithm. Simulations were validated by obtaining 5 million shotgun reads from four pregnant mothers and each father, collected under an IRB-approved protocol. Parental genotypes were obtained with a 290K SNP array (see FIG. 14).

In simulations, the MLE-based approach achieved 99.0% accuracy for fetal fractions as low as 9%, with reported confidence levels that corresponded well to the overall accuracy. We validated these results with four real samples, all of which gave correct calls with calculated confidence levels above 99%. In contrast, implementation of the published algorithm by Chiu et al. required a fetal fraction of 18% to achieve 99.0% accuracy, and only 87.8% accuracy was achieved with 9% fetal DNA.

Determining fetal fraction from parental genotypes in conjunction with an MLE-based approach achieves higher accuracy than published algorithms in predicting fetal fraction early in first and second trimesters. Furthermore, the methods disclosed herein yield a confidence metric that is crucial in determining the reliability of the results, especially at low fetal fractions where ploidy detection is more difficult. The published methods use a less accurate threshold method for calling ploidy based on a large set of disomy training data, an approach that predefines the false positive rate. Furthermore, without a confidence metric, there is a risk of reporting a false negative result when there is insufficient fetal cfDNA to make the call with the published method. In some embodiments, a confidence estimate is calculated for the called ploidy state.

Experiment 2
The objective was to improve non-invasive detection of fetal trisomies 18, 21, and X, especially in samples with low fetal fraction, by using a targeted sequencing approach in combination with parental genotype and Hapmap data in a Bayesian maximum likelihood estimation (MLE) algorithm.

Maternal and respective paternal samples from four euploid and two trisomy-positive pregnancies were obtained from patients with known fetal karyotypes under an IRB-approved protocol. Maternal cfDNA was extracted from plasma and after preferential enrichment for targeted specific SNPs, approximately 10 million sequence reads were obtained. Parental samples were similarly sequenced to obtain genotypes.

The described algorithm accurately called disomies 18 and 21 for all normal chromosomes in euploid and aneuploid samples. Calls of trisomies 18 and 21 were accurate, as were X-chromosome copy numbers in male and female fetuses. The confidence generated by the algorithm was greater than 98% in all cases.

The described method accurately reported ploidy for all tested chromosomes from six samples, including samples composed of less than 12% fetal DNA, which represents approximately 30% of first trimester and early second trimester samples. A crucial difference between the MLE algorithm and the published method is that the MLE algorithm leverages parental genotype and Hapmap data to improve accuracy and generate confidence metrics. At low fetal fractions, all methods have low accuracy; it is important to accurately identify samples that do not have sufficient fetal cfDNA to make reliable calls. While others have used Y-chromosome-specific probes to estimate fetal fractions in male fetuses, fetal fractions can be estimated for both sexes by simultaneously genotyping the parents. Another inherent limitation of the published method using untargeted shotgun sequencing is that the accuracy of ploidy calls varies between chromosomes due to differences in factors such as GC richness. The targeted sequencing approach is largely independent of such chromosome-wide variations, resulting in more consistent performance across chromosomes.

Experiment 3
The objective was to determine whether trisomies could be detected with high confidence in triploid fetuses using novel informatics to analyze SNP loci of free-floating fetal DNA in maternal plasma.

After ultrasound abnormality, 20 mL of blood was drawn from the pregnant patient. After centrifugation, maternal DNA was extracted from the buffy coat (DNEASY, QIAGEN) and cell-free DNA was extracted from the plasma (QIAAMP QIAGEN). In both DNA samples, targeted sequencing was applied to SNP loci on chromosomes 2, 21, and X. From the set of all possible ploidy states, the most likely hypothesis was selected by maximum likelihood Bayesian estimation. The method determines fetal DNA fraction, ploidy state, and an explicit confidence level in the ploidy determination. No assumptions are made regarding the ploidy of the reference chromosomes. The diagnosis uses a test statistic that is independent of the number of sequence reads, which is the current state of the art.

The method correctly diagnosed trisomies for chromosomes 2 and 21. The offspring rate was estimated at 11.9% [CI 11.7-12.1]. Fetuses were found to have one maternal copy and two paternal copies of chromosomes 2 and 21, effectively with a confidence level of 1 (error probability <10 ⁻³⁰ ). This was achieved using 92,600 and 258,100 reads for chromosomes 2 and 21, respectively.

This is the first demonstration of non-invasive prenatal diagnosis of a trisomic chromosome from maternal blood, as confirmed by metaphase karyotype, and the fetus is triploid. Existing methods of non-invasive diagnosis do not detect aneuploidy in this sample. Current methods rely on excess sequence reads in the trisomic chromosome compared to a disomic reference chromosome, but triploid fetuses do not have a disomic reference. Furthermore, existing methods do not provide a similarly reliable determination of ploidy using this proportion of fetal DNA and sequence read counts. Extending the approach to all 24 chromosomes is straightforward.

Experiment 4
The following protocol was used for 800-plex amplification using standard PCR (meaning no nesting was used) of DNA isolated from maternal plasma from euploid pregnancies, and genomic DNA from 21 triploid cell lines as well. Library preparation and amplification involved single-tube blunting followed by A-tailing. Adapter ligation was performed using the ligation kit found in the AGILENT SURESELECT kit, and PCR was run for 7 cycles. STA was then performed for 15 cycles using 800 different primer pairs targeting SNPs on chromosomes 2, 21 and X (95°C for 30 sec; 72°C for 1 min; 60°C for 4 min; 65°C for 1 min; 72°C for 30 sec). Reactions were run at a primer concentration of 12.5 nM. DNA was then sequenced using an ILLUMINA IIGAX sequencer. The sequencer output 1.9 million reads, 92% of which were mapped to the genome, and over 99% of the reads that were mapped to the genome were mapped to one of the regions targeted by the targeted primers. The numbers were essentially the same in both plasma and genomic DNA. Figure 15 shows the ratio of the two alleles for approximately 780 SNPs detected by the sequencer in genomic DNA obtained from a cell line with a known trisomy at chromosome 21. Note that the allele distribution is not easy to read visually, so the allele ratios are plotted here for ease of visualization. Circles indicate SNPs on disomic chromosomes, and stars indicate SNPs on trisomic chromosomes. Figure 16 is another display of the same data as in Figure X, where the Y-axis is the relative number of A and B measured for each SNP, and the X-axis is the number of SNPs split by chromosome. In Figure 16, SNPs 1-312 are found on chromosome 2, SNPs 313-605 are found on trisomic chromosome 21, and SNPs 606-800 are found on chromosome X. Data from chromosomes 2 and X show the disomic chromosomes as the relative sequence counts fall into three clusters: AA at the top of the graph, BB at the bottom of the graph, and AB in the center of the graph. Data from trisomic chromosome 21 shows four clusters: AAA at the top of the graph, AAB around the 0.65 line (2/3), ABB around the .35 line (1/3), and BBB at the bottom of the graph.

17A-D show data for the same 800plex protocol, measured on DNA amplified from four plasma samples from pregnant women. For these four samples, seven clusters of points are expected to be observed: (1) along the top of the graph are loci where the mother and fetus are both AA, (2) slightly below the top of the graph are loci where the mother is AA and the fetus is AB, (3) slightly above the 0.5 line are loci where the mother is AB and the fetus is AA, (4) along the 0.5 line are loci where the mother and fetus are both AB, (5) slightly below the 0.5 line are loci where the mother is AB and the fetus is BB, (6) slightly above the bottom of the graph are loci where the mother is BB and the fetus is AB, and (1) along the bottom of the graph are loci where the mother and fetus are both BB. The smaller the fetal fraction, the smaller the separation between clusters (1) and (2), between clusters (3), (4) and (5), and between clusters (6) and (7). The separation is expected to be half that of the fraction of DNA of fetal origin. For example, if 20% of the DNA is fetal and 80% is maternal, then (1)-(7) are expected to cluster at 1.0, 0.9, 0.6, 0.5, 0.4, 0.1 and 0.0, respectively; see e.g., FIG. 17D, POOL1_BC5_ref_rate. If instead 8% of the DNA is fetal and 92% is maternal, then (1)-(7) are expected to cluster at 1.00, 0.96, 0.54, 0.50, 0.46, 0.04 and 0.00, respectively; see e.g., FIG. 17B, POOL1_BC2_ref_rate. If no fetal DNA is detected, then (2), (3), (5), or (6) are not expected to be seen; alternatively, the separation can be said to be 0, so (1) and (2) are on top of each other, as are (3), (4) and (5), and similarly (6) and (7); see, for example, FIG. 17C, POOL1_BC7_ref_rate. Note that for FIG. 17A, POOL1_BC1_ref_rate, the fetal fraction is approximately 25%.

Experiment 5
Most methods of DNA amplification and measurement cause some allelic bias, where the two alleles found at a locus are generally detected with an intensity or count value that does not represent the actual amount of the allele in the DNA sample.For example, for a single individual, a 1:1 ratio of two alleles is expected to be observed at a heterozygous locus, which is the theoretical ratio expected for a heterozygous locus, but due to allelic bias, 55:45 or even 60:40 may be observed.In the context of sequencing, it is also noted that simple stochastic noise may result in significant allelic bias when the read depth is low.In an embodiment, it is possible to model the behavior of each SNP, so that if a consistent bias is observed for a particular allele, this bias can be corrected.Figure 18 shows the proportion of data that can be explained by binomial variance before and after bias correction. In Figure 18, stars indicate the allelic bias observed in the raw sequence data for the 800-plex experiment, and circles indicate the corrected allelic bias. Note that in the absence of any allelic bias, the data would be expected to fall along the x = y line. A similar set of data generated by amplifying DNA with 150-plex targeted amplification produced data that fell very close to the 1:1 line after bias correction.

Experiment 6
Universal amplification of DNA using adaptors ligated with primers specific to adaptor tags, with primer annealing and extension times limited to a few minutes, has the effect of enriching the proportion of shorter DNA strands. Most library protocols designed to generate DNA libraries suitable for sequencing include such steps, and examples of protocols are published and well known to those skilled in the art. In some embodiments of the present invention, adaptors with universal tags are ligated to plasma DNA and amplified using primers specific to adaptor tags. In some embodiments, the universal tag may be the same tag used for sequencing, it may be a universal tag only for PCR amplification, or it may be a collection of tags. Since fetal DNA is generally short in nature, while maternal DNA can be both short and long in nature, this method has the effect of enriching the proportion of fetal DNA in the mixture. The free-floating DNA, which is believed to be DNA from apoptotic cells and contains both fetal and maternal DNA, is short, mostly less than 200 bp. Cellular DNA released by cell lysis, a common phenomenon after phlebotomy, is generally almost exclusively maternal and is also quite long, mostly over 500 bp. Thus, blood samples left for more than a few minutes contain a mixture of short (fetal + maternal) and longer (maternal) DNA. By performing universal amplification with a relatively short extension time on maternal plasma followed by targeted amplification, the relative proportion of fetal DNA tends to increase compared to plasma amplified using targeted amplification alone. This can be seen in Figure 19, which shows the percent fetal measured when the input is plasma DNA (vertical axis) versus the percent fetal measured when the input DNA is plasma DNA with libraries prepared using the ILLUMINA GAIIx library preparation protocol. All points that fall below the line indicate that the library preparation step enriches the fraction of DNA of fetal origin. The two plasma samples that were red indicated hemolysis and therefore increased the amount of long maternal DNA present due to cell lysis, which indicates that the fetal fraction is particularly significantly enriched when library preparation is performed prior to targeted amplification. The methods disclosed herein are particularly useful when there is hemolysis or some other situation occurs where cells containing relatively long strands of contaminating DNA are lysed, contaminating the sample with mixed short and long DNA. Generally, the relatively short annealing and extension times are between 30 seconds and 2 minutes, but can be as short as 5 or 10 seconds or less, or as long as 5 or 10 minutes.

Experiment 7
The following protocols were used for direct PCR protocol and 1,200-plex amplification of DNA isolated from maternal plasma from euploid pregnancies, as well as genomic DNA from 21 triploid cell lines, as well as semi-nested approach. Library preparation and amplification involved single-tube blunting followed by A-tailing. Adapter ligation was performed using a modification of the ligation kit found in the AGILENT SURESELECT kit, and PCR was run for 7 cycles. For the targeted primer pool, 550 assays for SNPs from chromosome 21 and 325 assays for SNPs from each of chromosomes 1 and X were performed. Both protocols involved 15 cycles of STA using a primer concentration of 16 nM (95°C for 30 sec; 72°C for 1 min; 60°C for 4 min; 65°C for 30 sec; 72°C for 30 sec). The semi-nested PCR protocol involved a second amplification of STA15 cycles with an inner forward tag concentration of 29 nM and a reverse tag concentration of 1 μM or 0.1 μM (95°C for 30 sec; 72°C for 1 min; 60°C for 4 min; 65°C for 30 sec; 72°C for 30 sec). The DNA was then sequenced using an ILLUMINA II GAX sequencer. For the direct PCR protocol, 73% of the reads map to the genome, and for the semi-nested protocol, 97.2% of the sequence reads map to the genome. Thus, the semi-nested protocol provides approximately 30% more information, which is speculated to be mainly due to the elimination of primers that are most likely to cause primer dimers.

Variability in read depth tends to be higher using the semi-nested protocol than using the direct PCR protocol (see Figure 20), where the diamonds refer to read depth for loci run with the semi-nested protocol and the boxes refer to read depth for loci run without nesting. SNPs are arranged by read depth for the diamonds, so that the diamonds all lie on a curve, while the boxes appear to be loosely correlated; the placement of the SNPs is arbitrary, and it is the height of the dot, not the location of the dot from left to right, that refers to read depth.

In some embodiments, the methods described herein can achieve excellent depth of read (DOR) distribution. For example, in one version of this experiment (FIG. 21) using 1,200-plex direct PCR amplification of genomic DNA with 1,200 assays, 1186 assays had a DOR of over 10 and an average read depth of 400, and 1063 assays (88.6%) had a read depth between 200 and 800, an ideal window in which the number of reads for each allele was high enough to obtain meaningful data, but not so high that the number of reads for each allele was particularly small given the marginal use of these reads. Only 12 alleles had high read depth, with 1035 reads being the deepest. The standard deviation of DOR was 290, the average DOR was 453, the coefficient of variation of DOR was 64%, there were 950,000 total reads, and 63.1% of the reads were mapped to the genome. In another experiment (Figure 22) using a 1,200-plex semi-nested protocol, the DOR was higher. The standard deviation of the DOR was 583, the mean DOR was 630, the coefficient of variation of the DOR was 93%, there were 870,000 total reads, and 96.3% of the reads were mapped to the genome. Note that in both of these cases, the SNPs are arranged by the read depth for the mother, and therefore the curves show the maternal read depth. The differentiation between the child and the father is not significant, it is simply the trend that is important for this explanation.

Experiment 8
In the experiment, DNA from one cell and three cells were amplified using a semi-nested 1,200-plex PCR protocol. The experiment is in conjunction with prenatal aneuploidy testing using fetal cells isolated from maternal blood or for preimplantation genetic diagnosis using biopsy blastomere or trophectoderm samples. There were three replicates of one cell and three cells from two individuals (46XY and 47XX+21) per condition. The assay targeted chromosomes 1, 21 and X. Three different lysis methods were used: ARCTURUS, MPERv2 and alkaline lysis. Sequencing was performed on multiplexed 48 samples in one sequencing lane. The algorithm produced accurate ploidy calls for each of the three chromosomes and for each of the replicates.

Experiment 9
In one experiment, four maternal plasma samples were prepared and amplified using a hemi-nested 9,600-plex protocol. Samples were prepared as follows: up to 40 mL of maternal blood was centrifuged to isolate the buffy coat and plasma. Genomic DNA of the maternal samples was prepared from the buffy coat and paternal DNA was prepared from blood or saliva samples. Cell-free DNA in maternal plasma was isolated using the QIAGEN CIRCULATING NUCLEIC ACID kit and eluted in 45 μL of TE buffer according to the manufacturer's instructions. Universal ligation adaptors were added to the end of each molecule of 35 μL of purified plasma DNA and the library was amplified for seven cycles using adaptor-specific primers. The library was purified using AGENCOURT AMPURE beads and eluted in 50 μL of water.

3 μl of DNA was amplified using 15 cycles of STA with a primer concentration of 14.5 nM of 9,600 target-specific tagged reverse primers and 500 nM of one library adaptor-specific forward primer (initial polymerase activation at 95°C for 10 min, then 15 cycles of 95°C for 30 s; 72°C for 10 s; 65°C for 1 min; 60°C for 8 min; 65°C for 3 min and 72°C for 30 s; and a final extension at 72°C for 2 min).

The hemi-nested PCR protocol involved a second amplification of a dilution of the first STA product with a reverse tag concentration of 1000 nM and STA for 15 cycles using a concentration of 16.6 u nM for each of the 9,600 target-specific forward primers (initial 95°C for 10 min for polymerase activation, then 15 cycles of 95°C for 30 s; 65°C for 1 min; 60°C for 5 min; 65°C for 5 min and 72°C for 30 s; and a final extension at 72°C for 2 min).

Aliquots of the STA products were then amplified by standard PCR with 1 μM tag-specific forward primer and barcoded reverse primer for 10 cycles to generate barcoded sequencing libraries. Aliquots of each library were mixed with libraries of different barcodes and purified using spin columns.

Thus, 9,600 primers were used in a single-well reaction; primers were designed to target SNPs found on chromosomes 1, 2, 13, 18, 21, X and Y. The amplified products were then sequenced using an ILLUMINA GAIIX sequencer. Approximately 3.9 million reads were generated by the sequencer per sample, 3.7 million reads were mapped to the genome (94%), of which 2.9 million reads (74%) were mapped to the targeted SNPs, with an average read depth of 344 and a median read depth of 255. The fetal fractions for the four samples were found to be 9.9%, 18.9%, 16.3%, and 21.2%.

Related maternal and paternal genomic DNA samples were amplified and sequenced using a semi-nested 9600plex protocol. The semi-nested protocol differs in that 9,600 outer forward and tagged reverse primers are applied at 7.3 nM in the first STA. Thermocycling conditions and composition of the second STA, and barcoding PCR were the same as for the hemi-nested protocol.

The sequencing data was analyzed using the informatics methods disclosed herein to call the ploidy state for six chromosomes for fetuses whose DNA was present in the four maternal plasma samples. Ploidy calls for all 28 chromosomes in the population were called correctly with confidence levels greater than 99.2% except for one chromosome that was called correctly with a confidence level of 83%.

Figure 23 shows the read depth of the 9,600-plex hemi-nested approach along with the read depth of the 1,200-plex semi-nested approach described in experiment 7, where the number of SNPs with read depths >100, >200, and >400 were significantly higher than in the 1,200-plex protocol. The number of reads at the 90th percentile can be divided by the number of reads at the 10th percentile to obtain a dimensionless metric that indicates the uniformity of the read depth; the smaller the number, the more uniform (narrow) the read depth. The average 90th/10th percentile ratio is 11.5 for the method performed in experiment 9, but 5.6 for the method performed in experiment 7. Since fewer sequence reads are needed to ensure that a certain percentage of reads exceeds the read number threshold, a narrower read depth for a given protocol plexity is better for sequencing efficiency.

Experiment 10
In one experiment, four maternal plasma samples were prepared and amplified using a semi-nested 9,600-plex protocol. Details of experiment 10 were very similar to experiment 9, with the exception of the nesting protocol and the inclusion of four sample identities. Ploidy calls for all 28 chromosomes in the population were called correctly with confidence levels exceeding 99.7%. 7.6 million (97%) reads were mapped to the genome, and 6.3 million (80%) of the reads were mapped to targeted SNPs. The average read depth was 751, and the median read depth was 396.

Experiment 11
In one experiment, three maternal plasma samples were split into five equal parts, and each part was amplified using either 2,400 multiplex primers (four parts) or 1,200 multiplex primers (one part), for a total of 10,800 primers, using a semi-nested protocol. After amplification, the parts were pooled together for sequencing. The details of experiment 11 were very similar to experiment 9, with the exception of the nesting protocol and the split-and-pool approach. For all 21 chromosomes in the population, ploidy calls were called correctly, with confidence exceeding 99.7%, except for one that was not called well, with a confidence of 83%. 3.4 million reads were mapped to the targeted SNPs, with an average read depth of 404 and a median read depth of 258.

Experiment 12
In one experiment, four maternal plasma samples were divided into four equal parts, and each part was amplified using 2,400 multiplex primers, for a total of 9,600 primers, using a semi-nested protocol. After amplification, the parts were pooled together for sequencing. The details of experiment 12 were very similar to experiment 9, with the exception of the nesting protocol and the split-and-pool approach. Ploidy calls for all 28 chromosomes in the population were called correctly, with confidence levels exceeding 97%, except for one poorly called call with a confidence level of 78%. 4.5 million reads were mapped to the targeted SNPs, with an average read depth of 535 and a median read depth of 412.

Experiment 13
In one experiment, four maternal plasma samples were prepared and amplified using a 9,600-plex triple hemi-nested protocol for a total of 9,600 primers. The details of experiment 12 were very similar to experiment 9, with the exception of a nested protocol with three rounds of amplification; the three rounds involved 15 cycles of STA, 10 cycles of STA, and 15 cycles of STA, respectively. Ploidy calls for 27 of the 28 chromosomes in the population were called correctly with confidence exceeding 99.9%, except for one that was called correctly at 94.6% and one that was not called well with confidence of 80.8%. 3.5 million reads were mapped to the targeted SNPs, with an average read depth of 414 and a median read depth of 249.

Experiment 14 In one experiment, 45 sets of cells were amplified using a 1,200-plex semi-nested protocol, sequenced, and ploidy determinations were performed on three chromosomes. Note that this experiment is intended to simulate the conditions of performing preimplantation genetic diagnosis on single cell biopsies from day 3 embryos or trophectoderm biopsies from day 5 embryos. 15 individual single cells and 30 sets of three cells were placed into 45 individual reaction tubes for a total of 45 reactions, each reaction contained cells from only one cell line, but different reactions contained cells from different cell lines. Cells were prepared in 5 μl of wash buffer and lysed by adding 5 μl of ARCTURUS PICOPURE lysis buffer (APPLIED BIOSYSTEMS) and incubated at 56°C for 20 minutes and 95°C for 10 minutes.

Single cell/triple cell DNA was amplified with 25 cycles of STA using 1200 target-specific forward primers and tagged reverse primers at a primer concentration of 50 nM (initial polymerase activation at 95°C for 10 min, then 25 cycles of 95°C for 30 s; 72°C for 10 s; 65°C for 1 min; 60°C for 8 min; 65°C for 3 min and 72°C for 30 s; and a final extension at 72°C for 2 min).

The semi-nested PCR protocol involved three parallel second amplifications of dilutions of the first STA products over 20 cycles of STA (initial polymerase activation at 95°C for 10 min, then 95°C for 30 s; 65°C for 1 min; 60°C for 5 min; 65°C for 5 min and 72°C for 30 s; 15 cycles, and a final extension at 72°C for 2 min) using a reverse tag-specific primer at a concentration of 1000 nM and 400 target-specific nested forward primers at a concentration of 60 nM each. Thus, in three parallel 400-plex reactions, a total of 1200 targets amplified in the first STA were amplified.

Aliquots of the STA products were then amplified by standard PCR for 15 cycles with 1 μM tag-specific forward and barcoded reverse primers to generate barcoded sequencing libraries. Aliquots of each library were mixed with libraries of different barcodes and purified using spin columns.

Thus, 1,200 primers were used in the single-cell reaction; primers were designed to target SNPs found on chromosomes 1, 21, and X. The amplified products were then sequenced using an ILLUMINA GAIIX sequencer. Approximately 3.9 million reads per sample were generated by the sequencer, and 500-800 billion reads were mapped to the genome (74%-94% of all reads per sample).

Related maternal and paternal genomic DNA samples from cell lines were analyzed and sequenced using the same semi-nested 1200-plex assay pool, with a similar protocol, but with fewer cycles and a second STA of 1200-plex.

The sequencing data was analyzed using the informatics methods disclosed herein to call the ploidy state of the three chromosomes for the samples.

Figure 24 shows the normalized read depth ratios (vertical axis) for three chromosomes (1 = chromosome 1; 2 = chromosome 21; 3 = chromosome X) for six samples. The ratio was set equal to the number of reads mapping to that chromosome, normalized, and divided by the number of reads mapping to that chromosome averaged across three wells, each containing three 46XY cells. The set of three data points corresponding to a 46XY reaction is expected to have a 1:1 ratio. The set of three data points corresponding to a 47XX+21 cell is expected to have a 1:1 ratio for chromosome 1, 1.5:1 for chromosome 21, and 2:1 for chromosome X.

25 shows allelic ratios plotted for three chromosomes (1, 21, X) for three reactions. The bottom left reaction shows reactions in three 46XY cells. The left region is the allelic ratio for chromosome 1, the center region is the allelic ratio for chromosome 21, and the right region is the allelic ratio for the X chromosome. For 46XY cells, for chromosome 1, ratios of 1, 0.5, and 0 are expected to be observed, corresponding to SNP genotypes AA, AB, and BB. For 46XY cells, for chromosome 21, ratios of 1, 0.5, and 0 are expected to be observed, corresponding to SNP genotypes AA, AB, and BB. For 46XY cells, for the X chromosome, ratios of 1 and 0 are expected to be observed, corresponding to SNP genotypes A and B. The bottom right reaction shows reactions in three 47XX+21 cells. Allelic ratios are segregated by chromosome as in the bottom left graph. For 47XX+21 cells, for chromosome 1, ratios of 1, 0.5 and 0 are expected to be observed, corresponding to SNP genotypes AA, AB and BB. For 47XX+21 cells, for chromosome 21, ratios of 1, 0.67, 0.33 and 0 are expected to be observed, corresponding to SNP genotypes AAA, AAB, ABB and BBB. For 47XX+21 cells, for the X chromosome, ratios of 1, 0.5 and 0 are expected to be observed, corresponding to SNP genotypes AA, AB and BB. The plot in the top right was done for a reaction containing 1 ng of genomic DNA from the 47XX+21 cell line. Figure 26 shows the same graph as Figure 25, but for a reaction performed on only one cell. The graph on the left was for a reaction containing 47XX+21 cells, and the graph on the right was for a reaction containing 46XX cells.

From the graphs shown in Figures 25 and 26, it is visually apparent that there are two clusters of dots for chromosomes where a ratio of 1 and 0 is expected to be observed, three clusters of dots for chromosomes where a ratio of 1, 0.5, and 0 is expected to be observed, and four clusters of dots for chromosomes where a ratio of 1, 0.67, 0.33, and 0 is expected to be observed. The parental support algorithm was able to make accurate calls for all three chromosomes in all 45 reactions.

Experiment 15
In one experiment, maternal plasma samples were prepared and amplified using a hemi-nested 19,488-plex protocol. Samples were prepared as follows: up to 20 mL of maternal blood was centrifuged to isolate the buffy coat and plasma. Maternal genomic DNA was prepared from the buffy coat and paternal DNA was prepared from blood or saliva samples. Cell-free DNA in maternal plasma was isolated using the QIAGEN CIRCULATING NUCLEIC ACID kit and eluted in 50 μL of TE buffer according to the manufacturer's instructions. Universal ligation adaptors were added to the end of each molecule of 40 μL of purified plasma DNA and the library was amplified for nine cycles using adaptor-specific primers. The library was purified using AGENCOURT AMPURE beads and eluted in 50 μL of DNA suspension buffer.

6 μl of DNA was amplified with 15 cycles of STAR1 using primer concentrations of 7.5 nM of 19,488 target-specific tagged reverse primers and 500 nM of one library adaptor-specific forward primer (initial polymerase activation at 95°C for 10 min, then 15 cycles of 96°C for 30 s; 65°C for 1 min; 58°C for 6 min; 60°C for 8 min; 65°C for 4 min and 72°C for 30 s; and a final extension at 72°C for 2 min).

The hemi-nested PCR protocol involved a second amplification of a dilution of the first STAR1 product with a reverse tag concentration of 1000 nM and a concentration of 20 nM for each of the 19,488 target-specific forward primers for 15 cycles (STAR2) (initial polymerase activation at 95°C for 10 min, then 15 cycles of 95°C for 30 s; 65°C for 1 min; 60°C for 5 min; 65°C for 5 min and 72°C for 30 s; and a final extension at 72°C for 2 min).

Aliquots of the STAR2 products were then amplified by standard PCR for 12 cycles with 1 μM tag-specific forward and barcoded reverse primers to generate barcoded sequencing libraries. Aliquots of each library were mixed with libraries of different barcodes and purified using spin columns.

Thus, 19,488 primers were used in a single-well reaction. Primers were designed to target SNPs found on chromosomes 1, 2, 13, 18, 21, X, and Y. The amplified products were then sequenced using an ILLUMINA GAIIX sequencer. Approximately 10 million reads were generated by the sequencer for the plasma samples, 9.4-9.6 million reads were mapped to the genome (94-96%), of which 99.95% were mapped to the targeted SNPs, with an average read depth of 460 and a median read depth of 350. For comparison, a perfectly even distribution would be 10 million reads/19,488 targets = 513 reads/target. For primer dimers, 30,000 reads (0.3% of the reads generated by the sequencer) were from sequenced primer dimers. For genomic samples, 99.4-99.7% of reads were mapped to the genome, of which 99.99% were mapped to the targeted SNPs, and 0.1% of the sequencer-generated reads were primer dimers.

For a plasma sample with 10 million sequencing reads, typically at least 19,350 of the 19,488 target SNPs (99.3%) are amplified and sequenced. For a DNA sample with 2 million sequencing reads, typically at least 19,000 target SNPs (97.5%) are amplified and sequenced. The low numbers may be due to random sampling noise because the number of reads is small and the sequencer misses some amplification products. If necessary, the number of sequencing reads can be increased to increase the number of target SNPs that are amplified and sequenced.

Relevant maternal and paternal genomic DNA samples were amplified using 7.5 nM semi-nested 19,488 outer forward primers and tagged reverse primers in STAR1. Thermocycling conditions and composition of STAR2, and barcoding PCR were the same as for the hemi-nested protocol.

The average fetal fraction for the 407 samples was found to be 14.8%. The sequencing data was analyzed using the informatics methods disclosed herein to call the ploidy state of four fetal chromosomes (13, 18, 21, Y) for which DNA was present in 378 of the 407 maternal plasma samples, and chromosome X in 375 of the 407 maternal plasma samples. Ploidy calls for all 1,887 chromosomes in the population were called correctly with greater than 90% confidence. 1,882 of the 1,887 calls were greater than 95% confident, and 1,862 of the 1,887 calls were greater than 99% confident.

Similar control experiments were performed with the plasma PCR protocol using water instead of DNA extracted from plasma. In six such trials, 5-6% of the sequencing reads were primer dimers. The remaining sequencing reads were due to background noise. This experiment shows that a small amount of primer dimers are formed (without hybridizing to other primers and forming amplification primer dimers) even when there is no nucleic acid sample with a target locus for the primer to hybridize to.

Experiment 16
The following experiment illustrates a representative method for designing and selecting a primer library that can be used in any of the multiplex PCR methods of the present invention. The goal is to select primers from the initial candidate primer library that can be used to simultaneously amplify a large number of target loci (or a subset of target loci) in a single reaction. For the initial set of candidate target loci, it is not necessary to design or select primers for each target locus. It is preferable to design and select primers for as many desired target loci as possible.

Step 1
A set of candidate target loci (e.g., SNPs) was selected based on publicly available information regarding desirable parameters of the target loci, such as SNP frequency or SNP heterozygosity rate within a target population (worldwide web at ncbi.nlm.nih.gov/projects/SNP/; Sherry ST, Ward MH, Kholodov M, et al. dbSNP: the NCBI database of genetic variation. Nucleic Acids Res. 2001 Jan 1;29(1):308-11, each of which is incorporated herein by reference in its entirety). For each candidate locus, one or more PCR primer pairs were designed using the Primer3 program (worldwide web at primer3.sourceforge.net; libprimer3 release 2.2.3, which are incorporated herein by reference in their entireties). If there was no viable design for PCR primers for a particular target locus, that target locus was eliminated from further consideration.

Optionally, a "target locus score" (higher scores indicating higher desirability) may be calculated for most or all of the target loci, e.g., calculated based on a weighted average of parameters desirable for the various target loci. Parameters may be assigned different weights depending on their importance to the particular application for which the primers are used. Exemplary parameters include the heterozygosity rate of the target locus, the prevalence associated with the sequence (e.g., polymorphism) of the target locus, the disease penetrance associated with the sequence (e.g., polymorphism) of the target locus, the specificity of the candidate primers used to amplify the target locus, the size of the candidate primers used to amplify the target locus, and the size of the target amplicon.

Step 2
Thermodynamic interaction scores were calculated between each primer and all primers for all other target loci from step 1 (see, e.g., Alawi, H. T. & SantaLucia, J., Jr. (1998), "Thermodynamics of Internal C-T Mismatches in DNA", Nucleic Acids Res. 26, 2694-2701; Peyret, N., Seneviratne, P. A., Alawi, H. T. & SantaLucia, J., Jr. (1999), "Nearest-Neighbor Thermodynamics and NMR of DNA Sequences with "Nearest-Neighbor Thermodynamics of Internal A-C Mismatches in DNA: Sequence Dependence and pH Effects," Biochemistry 37, 9435-9444.; "Nearest-Neighbor Thermodynamics of Internal A-C Mismatches in DNA: Sequence Dependence and pH Effects," Biochemistry 37, 9435-9444.; "Nearest-Neighbor Thermodynamics of Internal A-C Mismatches in DNA: Sequence Dependence and pH Effects," Biochemistry 37, 9435-9444. Neighbor Thermodynamic Parameters for Internal G-A Mismatches in DNA", Biochemistry 37, 2170-2179; and Alawi, H. T. & SantaLucia, J. , Jr. (1997), "Thermodynamics and NMR of Internal G-T Mismatches in DNA", Biochemistry 36, 10581-10594; MultiPLX2.1 (see Kaplinski L, Andreson R, Puurand T, Remm M. MultiPLX: automatic grouping and evaluation of PCR primers. Bioinformatics. 2005 Apr 15;21(8):1701-2, which are incorporated by reference in their entireties). This step generates a 2D matrix of interaction scores. The interaction score predicted the likelihood of a primer dimer containing two interacting primers. The score was calculated as follows:
Interaction score = max (-deltaG_2, 0.8 * (-deltaG_1))
In the formula,
deltaG_2=Gibbs energy (the energy required to break a dimer) for a dimer that is extendable at both ends by PCR, i.e., the 3' end of each primer anneals to the other primer; and deltaG_1=Gibbs energy for a dimer that is extendable at at least one end by PCR.

Step 3:
For each target locus, if there is more than one primer pair design, one design is selected using the following method:
1. For each primer pair design at a locus, find the worst (best) interaction score for the two primers of that design and all primers from all designs to all other target loci.
2. Select the design with the best (lowest) score among the worst interaction scores.

Step 4
A graph was constructed such that each node represented one locus and its associated primer pair design (e.g., maximum clique problem). An edge was created between each pair of nodes. Each edge was assigned a weight equal to the worst (best) interaction score among the primers associated with the two nodes connected by the edge.

Step 5
Where necessary, for every pair of designs at two different target loci where one primer from one design and one primer from another design would anneal to overlapping target regions, additional edges were added between the nodes for the two designs. The weights of these edges were set equal to the highest weight assigned in step 4. Thus, step 5 prevents the library from including primers that anneal to overlapping target regions and thus would interfere with each other during the multiplex PCR reaction.

Step 6
The initial interaction score threshold was calculated as follows:
weight_threshold=max(edge_weight)-0.05*(max(edge_weight)-min(edge_weight))
In the formula,
max(edge_weight) is the maximum edge weight in the graph,
min(edge_weight) is the minimum edge weight in the graph.
The starting thresholds were set as follows:
max_weight_threshold=max(edge_weight)
min_weight_threshold=min(edge_weight)

Step 7
A new graph was constructed with only edges with weights above weight_threshold and with the same set of nodes as the graph in step 5. This step therefore ignores interactions with scores below weight_threshold.

Step 8
Nodes (and all edges connected to the removed nodes) were removed from the flag in step 7 until there were no remaining edges. Nodes were removed by iteratively applying the following steps for each:
1. Find the node with the highest angle (with the highest number of edges). If there is more than one, choose an arbitrary one.
2. Define a set of nodes consisting of the nodes selected above and all nodes connected to it (excluding nodes at angles less than those selected above).
3. Select the node from the set with the minimum target locus score (smaller scores represent less desirable) from step 1. Remove that node from the graph.

Step 9
If the number of nodes remaining in the graph satisfied the number of target genes required for the multiplex PCR pool (within an acceptable error), the method continued at step 10.

If there were too many or too few nodes remaining in the graph, a binary search was performed to determine what threshold produced the desired number of nodes remaining in the graph. If there were too many nodes in the graph, the weighted threshold bounds were adjusted as follows:
max_weight_threshold=weight_threshold. Otherwise (if there are too few nodes in the graph), the weight threshold boundary was adjusted as follows:
min_weighted_threshold=weighted_threshold Then, the weighted threshold was adjusted as follows:
Weight_threshold=(max_weight_threshold+min_weight_threshold)/2
Steps 7-9 were repeated.

Step 10
Primer pair designs associated with the remaining nodes in the graph were selected for a primer library that can be used in any of the methods of the invention.

Optionally, this primer design and selection method can also be performed on primer libraries where only one primer (rather than a primer pair) is used to amplify a target locus. In this case, the nodes represent the target locus per primer (rather than a primer pair).

Experiment 17
27 is a graph comparing two primer libraries designed using the methods of the invention. The graph shows the number of loci with a particular minor allele frequency targeted by each primer library. More primers were retained during selection of the "New Pool" library. This library allows for the amplification of more target loci, particularly more target loci with relatively large minor allele frequencies (which are alleles that are more informative for some methods of the invention, e.g., detection of fetal chromosomal abnormalities).

These primer libraries were used in the following multiplex PCR method. Blood (20-40 mL) was collected from each subject into 2-4 CELL-FREE™ DNA tubes (Streck). Plasma (minimum 7 mL) was isolated from each sample using a double centrifugation protocol at 2,000 g for 20 min followed by 3,220 g for 30 min, with the supernatant decanted after the first centrifugation. cfDNA was isolated from 7-20 mL of plasma using the QIAGEN QIAamp Circulating Nucleic Acid kit and eluted with 45 uL of TE buffer. Pure maternal genomic DNA was isolated from the buffy coat obtained after the first centrifugation, and pure paternal genomic DNA was prepared similarly from blood, saliva or buccal samples.

Using the 11,000 target-specific assay, maternal cfDNA, maternal genomic DNA, and paternal genomic DNA samples were preamplified for 15 cycles and aliquots were transferred to a second PCR reaction with nested primers for 15 cycles. Finally, samples were prepared for sequencing by adding barcoded tags in a third 12-cycle round of PCR. Thus, 11,000 targets were amplified per reaction, including SNPs found on chromosomes 13, 18, 21, X, and Y. Amplification products were then sequenced using an ILLUMINA GAIIx or HISEQ sequencer. Parental genotypes were sequenced at a shallower read depth (approximately 20% of the cfDNA read depth) than fetal genotypes.

Experiment 18
If desired, PCR products can be analyzed for size and quantity using standard methods, for example, an Agilent Technologies 2100 Bioanalyzer (Figures 28A-M). For example, the non-nested direct PCR method described herein was used for the 2,400-plex (Figures 28B-28G) and 19,488-plex experiments (Figures 28H-28M). Primer amounts for Figures 28B-28D and 28H-28J were 10 nM. Primer amounts for Figures 28E-28G and 28K-28M were 1 nM. For Figures 28B, 28E, 28H, and 28K, the amount of input DNA was 24 ng; for Figures 28C, 28F, 28I, and 28L, 80 ng; and for Figures 28D, 28G, 28J, and 28M, 250 ng. More input DNA produced a greater proportion of the desired 180 base pair product. The peak at 140 base pair position is the primer dimer product.

Experiment 19
Proof-of-principle studies demonstrated detection of T13, T18, T21, 45,X, and 47,XXY with equally high accuracy across all chromosomes.

Patients Pregnant couples were enrolled in a specific prenatal care center under a protocol approved by the Institutional Review Board in accordance with local law. Inclusion criteria were age at least 18 years, gestational age at least 9 weeks, singleton pregnancy, and signed informed consent. Blood samples were taken from pregnant mothers and blood or buccal samples were collected from fathers. Samples from 2 pregnancies with T13 (Patau syndrome), 2 with T18 (Edwards syndrome), 2 with T21 (Down syndrome), 2 with 45,X, 2 with 47,XXY, and 90 normal pregnancies were selected from a cohort of approximately 500 women prior to testing to test which chromosomal abnormalities the method detects. When postnatal child tissue was available, normal fetal karyotype was confirmed by molecular karyotyping. Euploid samples were collected prior to invasive testing in low-risk women. Euploid samples were collected at least 7 days after invasive testing and aneuploidy was confirmed by cytogenetic karyotype analysis or fluorescent in situ hybridization in a separate laboratory.

Sample preparation and multiplex PCR
For the data in Figures 30A-E, 30G, 30H, and 31A-31G, sample preparation and 19,488-plex PCR were performed as described in experiment 15. For the data in Figure 30F, sample preparation and 11,000-plex PCR were performed as described in experiment 17.

Methods and Data Analysis The algorithm considers parental genotypes and crossover frequency data (e.g., from the HapMap database) to calculate a large number of possible fetal ploidy states and predicted allele distributions at 19,488 polymorphic loci for different fetal cfDNA fractions (Figures 29A-29C). Unlike allele ratio-based methods, the algorithm also takes into account linkage disequilibrium and uses a non-Gaussian data model to show predicted distributions of allele measurements for SNPs with observed platform characteristics and amplification bias. The algorithm then compares the different predicted allele distributions with the actual allele distributions measured in the cfDNA samples (Figure 29C) and calculates the likelihood of each hypothesis (monosomy, disomy, or trisomy, for which there are many hypotheses based on the different possible crossovers) based on the sequencing data. The algorithm sums the likelihood of each individual monosomy, disomy, or trisomy hypothesis (Figure 29D) and calls the ploidy state with the greatest total likelihood as copy number and fetal fraction (Figure 29E). Although laboratory investigators were not blinded to sample karyotype, the algorithm calls ploidy states without human intervention, thus remaining blind to the truth.

Interpretation of the Data Graphical Representation of Generated Data To determine the ploidy state of the chromosome of interest, the algorithm considers the distribution of sequence counts of the two possible alleles at each of the 3,000-4,000 SNP positions per chromosome. It is important to note that the algorithm makes ploidy calls using an approach that does not lend itself to visualization. Thus, for purposes of illustration, the data is presented herein in a simplified manner as the ratio of the two most likely alleles, labeled A and B, so that the associated trends can be more easily visualized. This simplified illustration does not take into account some features of the algorithm. For example, two important aspects of the algorithm that cannot be accounted for in the visualization method of presenting allele ratios are 1) the ability to exploit linkage disequilibrium, i.e., the impact of measurements at one SNP on the possible unique properties of adjacent SNPs, and 2) the use of a non-Gaussian data model that accounts for the distribution of expected allele measurements for SNPs with platform characteristics and amplification bias. Also note that the algorithm only considers the two most frequent alleles at each SNP position and ignores other possible alleles.

The graphical representations in Figures 30A-30H include samples in which two, one, or three fetal chromosomes are present. Typically, these indicate euploidy (Figures 30A-30C), monosomy (Figure 30D), and trisomy (Figures 30E-30H), respectively. In all plots, each spot represents a single SNP, and the target SNPs are plotted sequentially along the horizontal axis from left to right for one chromosome. The vertical axis represents the number of reads for the A allele as a ratio to the total number of reads for both A and B alleles for that SNP. Note that measurements were made on total cfDNA isolated from maternal blood, which includes both maternal and fetal cfDNA, and therefore each spot represents the combined contribution of fetal and maternal DNA to that SNP. Thus, an increase in the proportion of maternal cfDNA from 0% to 100% would cause some spots to gradually move up and down in the plot depending on maternal and fetal genotypes.

If needed to facilitate visualization, the spots can be colored according to maternal genotype, which is useful for visualizing ploidy status, since the maternal genotype contributes a lot to the localization of each spot and most trisomies are maternally inherited. Specifically, SNPs with maternal genotype AA can be shown in red, SNPs with maternal genotype AB can be shown in green, and SNPs with maternal genotype BB can be shown in blue.

In all cases, SNPs that are homozygous for the A allele (AA) from both the mother and fetus are seen to stick to the upper limit of the plot because the A allele read ratio is high since the B allele should not be present. Conversely, SNPs that are homozygous for the B allele from both the mother and fetus are seen to stick to the lower limit of the plot because the A allele read ratio is small since only the B allele should be present. Spots that do not stick to either the upper or lower limit of the plot indicate SNPs where the mother, fetus, or both are heterozygous. These spots are useful for identifying fetal ploidy, but may also be informative for determining paternal versus maternal inheritance. These spots separate based on both maternal and fetal genotypes and fetal fraction, and therefore the precise location of individual spots along each y-axis depends on both stoichiometry and fetal fraction. For example, loci where the mother is AA and the fetus is AB are expected to have different A allele read ratios and therefore occupy different positions along the y-axis depending on the fetal fraction.

Presence of Two Chromosomes Figures 30A-30C provide data showing the presence of two chromosomes when samples are entirely maternal (no fetal cfDNA present: Figure 30A), contain a medium fetal cfDNA fraction (Figure 30B), or contain a high fetal cfDNA fraction (Figure 30C).

Figure 30A shows data from cfDNA isolated from non-pregnant blood. When fetal cfDNA is absent and the sample contains only maternal cfDNA, the plot represents purely euploid maternal genotypes, with the hallmark pattern including "clusters" of spots, with a red cluster tightly attached to the top of the plot (SNPs with maternal genotype AA), a blue cluster tightly attached to the bottom of the plot (SNPs with maternal genotype BB), and a single green cluster in the middle (SNPs with maternal genotype AB) (color not shown in figure).

If fetal cfDNA is present, the locations of the spots shift such that the clusters separate into discrete "bands." Note that for samples with 0% fetal fraction, the cluster of spots is called a "cluster" (as in FIG. 30A), and for all samples with fetal fractions greater than 0%, the cluster of spots is called a "band" (as in FIGS. 30B-30J). If the fetal fraction is large enough, these discrete bands will be easily visible. Specifically, FIGS. 30B and 30C show a characteristic pattern associated with two fetal chromosomes with medium and high fetal fractions, respectively. This pattern includes three central green bands corresponding to SNPs that are heterozygous in the mother, and two "peripheral" bands at both the top (red) and bottom (blue) ends of the plot, respectively, corresponding to SNPs that are homozygous in the mother (color not shown in the figure).

Figure 30B shows data from cfDNA isolated from a plasma sample from a woman with a euploid fetus and containing a 12% fetal cfDNA fraction, where the cluster of spots that are tightly attached to the top and bottom of the plot, respectively, separate into two discrete bands (colors not shown in the figure): one red and one blue outer peripheral band that remain tightly attached to the top or bottom of the plot, and one red and one blue inner peripheral band that are away from the edge of the plot. These inner peripheral bands centered near 0.92 and 0.08 represent SNPs with maternal genotype AA and fetal genotype AB (shown in red), and SNPs with maternal genotype BB and fetal genotype AB (shown in blue), respectively. The central cluster of green spots is spread out, but separation into distinct bands is not readily visible in this fetal fraction.

At high fetal cfDNA fractions, the typical pattern indicating the presence of two chromosomes (three sets of green bands, and two red and two blue peripheral bands) is easily visible (color not shown in figure). Figure 30C shows data from a plasma sample from a woman carrying a euploid fetus at a fetal cfDNA fraction of 26%. Here, the peripheral bands move toward the center of the plot with the inner bands separating due to the altered B allele levels from the increased fetal cfDNA fraction. Significantly, at high fetal fractions, the separation of the central green cluster into three separate bands is now quite easily visible. This central triad of bands, in this case clusters near 0.37, 0.50, and 0.63, corresponding to SNPs with maternal genotype AB and fetal genotypes BB (bottom), AB (middle), and AA (top), respectively.

These hallmark patterns, i.e., three green bands and four peripheral bands (two red and two blue), indicate the presence of two chromosomes, as in the case of autosomal euploidy, or an X chromosome in a female (XX) fetus.

Presence of One Chromosome Fetal heterozygosity is not possible if the fetus inherits only a single chromosome and therefore only a single allele. Thus, the only possible fetal SNP identity is A or B. Thus, maternally inherited monosomic chromosomes have a characteristic pattern of two central green bands indicating SNPs for which the mother is heterozygous, and unique red and blue peripheral bands (1 and 0 read ratios) indicating SNPs for which the mother is homozygous, respectively, remaining tightly attached to the top and bottom of the plot (Figure 30D) (color not shown in figure). Note the absence of inner peripheral bands. This pattern indicates the presence of one chromosome, as in the case of maternally inherited autosomal monosomy, or the X chromosome in a male (XY) fetus.

Presence of Three Chromosomes Trisomic chromosomes have three characteristic patterns. The first pattern indicates a maternally inherited meiotic trisomy (meiotic error), where the fetus inherits two homologous, non-identical chromosomes from the mother (Figure 30E). This pattern includes two central green bands with two red and blue peripheral bands, respectively (color not shown in figure). The second pattern indicates a paternally inherited meiotic trisomy, where the fetus inherits two homologous, non-identical paternally derived chromosomes (Figure 30F). This pattern includes four central green bands and three red and blue peripheral bands, respectively (color not shown in figure). The third pattern indicates a maternally (Figure 30G) or paternally inherited (Figure 30H) mitotic trisomy (mitotic error), where the fetus inherits two identical maternal or paternal chromosomes. This pattern includes four central green bands with two red and blue peripheral bands, respectively. Maternally and paternally inherited mitotic trisomies can be distinguished by adjacent arrangement of red and blue bands. In paternally inherited mitotic trisomies, the red and blue inner peripheral bands (bands that are not attached to the edge of the plot) are closer to the center (not shown in the figure). This is due to the paternal contribution of the same chromosome. Note that our previous results indicate that at the blastomere stage, 66.7% of maternally inherited trisomies are meiotic and only 10.2% of trisomies are paternally inherited.

For the Y chromosome, the PS method considers a separate set of hypotheses: the presence of zero, one, or two chromosomes. Since heterozygous loci are not possible in the absence of maternal contribution to the sequence reads at each locus (the two Y chromosomes case necessarily involves two identical chromosomes), the bands remain tightly attached to the top (A allele) or bottom (B allele) of the plot (data not shown), making the analysis extremely simple and dependent on quantitative allele count data. Note that since the method interrogates SNPs, it uses homologous non-recombining SNPs from the Y chromosome, thus obtaining data on both X and Y for one probe pair.

Identifying Aneuploidies As mentioned above, identifying autosomal aneuploidies using this plot-based visualization method is straightforward given sufficient fetal fraction and only requires identifying plots where abnormal numbers of chromosomes are present. By summing the copy number information of X and Y chromosomes, the presence or absence of sex chromosome aneuploidies is identified. Specifically, a plot representing a fetus with a 47,XXX genotype has a typical "three chromosome" pattern, and a plot representing a fetus with a 47,XXY genotype has a typical "two chromosome" pattern for the X chromosome, but also has allele reads indicating the presence of one Y chromosome. Similarly, the method can call 47,XYY, where the "one chromosome" pattern indicates the presence of a single X chromosome and the allele reads indicate the presence of two Y chromosomes. A fetus with a 45,X genotype has a typical "one chromosome" pattern for the X chromosome, and the data indicates zero Y chromosomes.

Effect of fetal fraction As discussed above, the number of fetal sequence reads gives the exact location of each spot along the y-axis of the plot. The fetal fraction affects the proportion of fetal and maternal reads, and therefore the positioning of each spot. At a high percentage of fetal cfDNA (usually around 20%) in Figures 30C-30E and Figures 30G and 30H, it is easily seen that the spot clusters are primarily based on the maternal genotype, but the presence of fetal DNA from alleles whose genotype differs from the maternal genotype changes the cluster into multiple distinct bands. However, as the fetal fraction decreases (as in Figures 30B and 30F), the spots move back towards the ends and center of the plot, resulting in tighter clusters. Specifically, the set of red peripheral bands with maternal genotype AA will move back towards the top of the plot, the set of blue peripheral bands with maternal genotype BB will move back towards the bottom, and the set of green central bands with heterozygous mothers will be compressed into a single cluster in the center of the plot (compare Figures 30B and 30C) (color not shown in figure). Although aneuploidy is not easily visible using this visualization technique for low fetal fractions, the algorithm can identify ploidy states for very small fetal fractions, e.g., 3% fetal fraction. This is achieved by using statistical techniques to compare observed data to highly accurate data models that predict allele distributions for a given set of sample parameters (e.g., copy number, parental genotypes, and fetal fraction, etc.). For small fetal fractions, data model accuracy is important because the difference between allele distributions for different ploidy states is proportional to the fetal fraction. Furthermore, the algorithm can determine when a data set does not contain enough data to make a reliable fetal ploidy determination.

Results The mapped sequencing reads of the targeted SNPs were deemed informative and used in the algorithm. More than 95% of the targeted loci were observed in the sequencing results. Plots visualizing the significant ploidy calls are shown in Figures 31A-31G. Figure 31A shows a euploid sample. Here, chromosomes 13, 18, and 21 have the typical "2-chromosome" pattern (as described herein). It contains three sets of green central bands, and two red and two blue peripheral bands. This, together with the presence of two green central bands for the X chromosome and a Y chromosome band along the edge of the plot, indicates a euploid XY genotype (color not shown in the figure).

The most common autosomal trisomies, T13, T18, and T21, are shown in the plots in Figures 31B, 31C, and 31D, respectively. Specifically, Figure 31B shows a T13 sample, where chromosomes 18 and 21 show the typical "two chromosome" pattern, chromosome X shows the typical "one chromosome" pattern, and there are reads from the Y chromosome. Overall, this shows disomy at chromosomes 18 and 21, identifying the XY genotype of the fetus. However, specifically, chromosome 13 shows the typical "three chromosome" pattern. Similarly, Figure 31C shows a T18 sample, and Figure 31D shows a T21 sample.

The method can also detect sex chromosome aneuploidies including 45,X (Figure 31E), 47,XXY (Figure 31F), and 47,XYY (Figure 31G). Note that the method calls copy numbers at chromosomes 13, 18, 21, X, and Y, and the overall chromosome count is reported assuming disomy for the remaining chromosomes. The X chromosome region of the plot showing the 45,X sample indicates the presence of a single chromosome. However, the lack of reads from the Y chromosome, in addition to the "two chromosome" pattern for chromosomes 13, 18, and 21, points to a 45,X genotype. Conversely, the 47,XXY sample produces a plot indicating the presence of two X chromosomes. The data also showed allelic reads from the Y chromosome. In addition to the two copies of chromosomes 13, 18, and 21, this points to a 47,XXY genotype. The 47,XYY genotype is indicated by the presence of a "one chromosome" pattern for the X chromosome, and a lead representing the presence of two Y chromosomes.

Discussion The method non-invasively detects T13, T18, T21, 45,X, 47,XXY, and 47,XYY from maternal blood. The method interrogates cfDNA from maternal plasma by targeted multiplex PCR amplification and high-throughput sequencing of 19,488 SNPs. The method, combined with the advanced informatics analysis of the method that takes into account parental genotype information and many sample parameters, including fetal fraction and DNA quality, more reliably detects fetal precursors and makes highly accurate ploidy calls on all five chromosomes associated with the seven most common types of birth aneuploidy (T13, T18, T21, 45,X, 47,XXX, 47,XXY, and 47,XYY). The method offers many clinical advantages over previous methods, including significantly greater clinical coverage and higher sample-specific computational accuracy (as well as individualized risk scores).

Expanded clinical scope Given its ability to accurately detect autosomal trisomies and sex chromosome aneuploidies, the method increases aneuploidy coverage by approximately 2-fold compared to clinically available NIPT methods. The method presented herein is the only non-invasive test that calls sex chromosome ploidy with high accuracy. Previous DNA mixing experiments and other plasma samples analyzed in our experimental assay suggest that the method can detect a larger cohort of sex chromosome abnormalities, including 47,XXX. It is also expected that the method presented herein can detect aneuploidy of chromosomes 13, 18, and 21 with high sensitivity and specificity, and with appropriate primer design, it can detect copy number of the remaining chromosomes as well.

Calculation of sample-specific accuracy It is particularly meaningful that the method calculates a sample-specific accuracy for the ploidy call of each chromosome in each sample. The accuracy calculated by the method is expected to greatly reduce the rate of incorrect calls by identifying and marking individual samples with poor quality DNA or low fetal fraction that may produce low accuracy test results. In contrast, methods based on large-scale shotgun sequencing (MPSS) make positive or negative calls using single hypothesis rejection tests, and their accuracy estimates are based on published research cohorts rather than on the characteristics of individual samples, and are assumed to have the same accuracy as the cohort. However, individual accuracy for samples with parameters in the tail of the cohort distribution can be significantly different. This phenomenon is exacerbated with low fetal fractions, such as in the case of early gestational age, or for samples with low DNA quality. These samples are usually not identified and tagged for follow-up observation, which may result in missed calls. However, the present invention takes into account many parameters, such as fetal fraction and many DNA quality measures, to make each chromosome copy number call and calculates a sample-specific accuracy for that call. This allows the method to identify and flag individual samples with low accuracy for follow-up. This is expected to eliminate most missed calls, especially in the early stages of pregnancy when the fetal fraction is usually low. The premise is that no calls are far preferable to missed calls, since a missed call can simply be redrawn and reanalyzed.

Conversion of Calculated Accuracy to Traditional Risk Score The method can provide an adjusted risk of aneuploidy for high-risk pregnant women, where the adjusted risk takes into account the prior risk (Benn P, Cuckle H, Pergament E. Non-invasive prenatal diagnosis for Down syndrome: the paradigm will shift, but slowly. Ultrasound Obstet Gynecol 2012;39:127-130, which is incorporated herein by reference in its entirety). Although the method provides calculated accuracy tailored to each patient, for clinical use these accuracy can be converted to a traditional risk score, which also represents the risk of euploid pregnancy, but expressed as a percentage. Conventional risk scores take into account various parameters such as maternal age-related risk and serum levels of biochemical markers to provide a risk score above which a mother is considered at high risk and is recommended to undergo invasive diagnostic procedures for follow-up. The method significantly improves the accuracy of this risk score, thus reducing both false positive and false negative rates and providing a more accurate assessment of individual maternal risk. The calculation accuracy used herein is the likelihood that the ploidy call is correct, expressed as a percentage, but the calculation accuracy used in experiment 19 does not include age-related risk. Since risk score calculations usually include age-related risk, the calculation accuracy and conventional risk scores are not interchangeable and need to be combined to convert to conventional risk scores. The formula for combining age-related risk with calculation accuracy is:
where _R1 is the risk score calculated by the method and _R2 is the risk score calculated by first trimester screening.

SNP-based methods eliminate the problem of amplification variation An inherent drawback of the counting method used by some other methods is that the fetal ploidy status is determined by measuring the ratio of the number of reads mapped to the target chromosome (e.g., chromosome 21) to the number of reads mapped to the reference chromosome. There is a large variation in the amplification of chromosomes with high or low GC content, including chromosomes 13, X, and Y. This can cause signal variations as large as the fetal cfDNA signal, which can lead to misleading copy number calls by changing the ratio of allele reads from the target chromosome to the reads from the reference chromosome. This can result in low accuracy for chromosomes 13, X, and Y. Importantly, this problem is exacerbated in the case of low fetal cfDNA percentages, which are likely to occur in early gestational age.

In contrast, SNP-based methods do not rely on consistent amplification levels between chromosomes and are therefore expected to give results of equal accuracy across chromosomes. Because the methods examine the relative numbers of alternative alleles at polymorphic loci that by definition differ by only a single nucleotide, the methods do not require the use of a reference chromosome, thereby avoiding problems with inter-chromosomal amplification variation inherent in methods that rely on quantification read numbers. Unlike quantitative methods that require a euploid reference chromosome, the methods are expected to be able to detect abnormalities that do not involve copy number changes, such as triploidy, as well as uniparental disomy.

Importance of Early Detection Significantly, the combined birth prevalence of sex chromosome aneuploidies is higher than that of the most common autosomal aneuploidies (Figure 32). However, there are currently no routine non-invasive screening methods that can reliably detect sex chromosome abnormalities. Thus, sex chromosome abnormalities are usually detected prenatally as a by-product of routine testing for Down's syndrome or other autosomal aneuploidies, with the majority of cases being missed entirely. Early and accurate detection is crucial for many of these disorders, where early intervention improves clinical outcomes. For example, the overall combined birth prevalence of Turner syndrome is 1 in 2,500 females, but Turner syndrome is often not diagnosed until adolescence. Growth hormone therapy is known to prevent the short stature caused by the disorder, but treatment is significantly more effective if initiated before age 4. Additionally, estrogen replacement therapy can stimulate secondary sexual characteristics in patients with Turner syndrome, but again, treatment should be initiated just before puberty, before the syndrome is routinely detected. All these highlight the importance of early, routine and safe detection of sex chromosome aneuploidies. The method provides the first potentially useful approach for routine screening for sex chromosome abnormalities.

Additional Applications The method is uniquely poised to detect submicroscopic abnormalities, such as microdeletions and microduplications, due to the use of targeted amplification. Non-targeted methods such as MPSS have been shown to be able to detect DiGeorge microdeletion syndrome, but the method requires a sufficiently high level of gene coverage to make the procedure difficult to implement.
This is why untargeted amplification is several orders of magnitude less efficient for submicroscopic regions, where only a very small proportion of sequencing reads are likely to be informative. Furthermore, the fact that currently available methods have problems accurately identifying ploidy states for sex chromosomes suggests that for smaller chromosomal segments, they face variable amplification problems as well.

Similarly, SNP-based methods can detect UPD disorders, which are non-copy number alteration abnormalities that cannot be detected by current non-invasive methods that rely on counting or traditional invasive methods such as amniocentesis and CVS that rely on cytogenetic karyotyping and/or fluorescent in situ hybridization. This is because clinically available MPSS-based targeting methods amplify non-polymorphic loci and therefore cannot determine, for example, whether the target chromosomes are from the same parent, while SNP-based methods can specifically identify individual haplotypes. This means that these microdeletions/microduplications and UPD syndromes, such as Prader-Willi, Angelman, and Beckwith-Wiedemann syndromes, are usually not diagnosed prenatally and are often misdiagnosed early after birth. This results in significant delays in therapeutic intervention. Moreover, because the method targets SNPs, it also facilitates reconstruction of parental haplotypes and allows detection of fetal inheritance of individual disease-linked loci (Kitzman JO, Snyder MW, Ventura M, et al. Noninvasive whole-genome sequencing of a human fetus. Sci Transl Med 2012;4:137ra76, which is incorporated herein by reference in its entirety).

The results presented herein demonstrate that the scope of the method can be expanded for prenatal aneuploidy identification. Specifically, by amplifying and sequencing 19,488 SNPs, the method can determine copy number of chromosomes 13, 18, 21, X, and Y, and is expected to specifically detect other chromosomal abnormalities, such as triploidy and UPD, that are not detected by any other clinically available non-invasive method. The increased clinical coverage and strong sample-specific computational accuracy suggest that the method can provide a useful adjunct to invasive testing for fetal chromosomal aneuploidy detection.

All patents, patent applications, and published references cited herein are incorporated herein by reference in their entirety. While the disclosed method has been described with certain embodiments thereof, it will be understood that it may be further modified. Furthermore, this application is intended to cover any variations, uses, or adaptations of the disclosed method, including departures from the present disclosure that are within known or customary practice in the art with respect to the disclosed method, and that are within the scope of the appended claims. For example, any of the methods disclosed herein for DNA can be readily adapted to RNA by introducing a reverse transcription step to convert the RNA to DNA. The examples using polymorphic loci for illustration can be readily adapted to the amplification of non-polymorphic loci, if desired.

Claims (12)

1. A method for preparing a deoxyribonucleic acid (DNA) fraction from a cancer patient useful for analyzing copy number polymorphisms and mutations in cancer in a subject, comprising:
(a ) extracting cell-free DNA from blood collected from the subject ;
( b ) (i) generating barcoded DNA by ligating adapter tags and molecular barcodes to the extracted cell-free DNA, wherein cell-free DNA molecules containing the same target locus are each tagged with a different molecular barcode and are distinguishable from one another by sequencing of the molecular barcodes;
(ii) generating a sequencing library from the barcoded DNA by performing universal amplification using the adaptor tags;
(iii) generating a fraction from the DNA extracted in (a) by enriching the plurality of loci from the sequencing library using hybrid capture probes targeting the plurality of loci, the plurality of loci comprising 100-2,000 loci; and ( c ) ( i ) performing massively parallel sequencing on the enriched plurality of loci to obtain sequencing reads for the plurality of loci.
(ii) analyzing the cell-free DNA in the fraction of DNA generated in ( b ) by determining the genetic status of the plurality of loci based on the sequencing reads obtained by the massively parallel sequencing to detect copy number variations and mutations of the cancer in the subject;
wherein the possible forms of genetic status include repeats and mutations. The method of claim 1, wherein the plurality of loci includes 100 to 1,000 loci. The method of claim 1, wherein the plurality of loci includes 300 to 2,000 loci. The method of claim 1, wherein the cell-free DNA comprises a mixture of DNA from the cancer and DNA from healthy tissue of the cancer patient. The method of claim 4, further comprising determining a cancer-derived DNA fraction based on the sequencing reads derived from the cancer DNA and DNA derived from healthy tissue of the cancer patient. 1. A method for preparing a deoxyribonucleic acid (DNA) fraction from a cancer patient useful for analyzing the ploidy status of cancer in a subject, comprising:
(a ) extracting cell-free DNA from blood collected from the subject ;
( b ) (i) generating barcoded DNA by ligating adapter tags and molecular barcodes to the extracted cell-free DNA, wherein cell-free DNA molecules containing the same target locus are each tagged with a different molecular barcode and are distinguishable from one another by sequencing of the molecular barcodes;
(ii) generating a sequencing library from the barcoded DNA by performing universal amplification using the adaptor tags;
(iii) generating a fraction from the DNA extracted in (a) by enriching the plurality of loci from the sequencing library using hybrid capture probes targeting the plurality of loci, the plurality of loci comprising 100-2,000 loci; and ( c ) ( i ) performing massively parallel sequencing on the enriched plurality of loci to obtain sequencing reads for the plurality of loci.
(ii) determining a ploidy status of the cancer in the subject based on the sequencing reads obtained by the massively parallel sequencing;
wherein the plurality of loci comprises between 300 and 2,000 loci. The method of claim 6, wherein the plurality of loci includes 300 to 1,000 loci. The method of claim 6, further comprising determining mutations at the plurality of loci based on the sequencing reads. The method of claim 8, further comprising determining the number of unique molecules in the blood sample for each locus based on the barcodes and sequencing reads from the cell-free DNA. The method of claim 6, wherein the cell-free DNA comprises a mixture of DNA from the cancer and DNA from healthy tissue of the cancer patient. The method of claim 10, further comprising determining a DNA fraction derived from the cancer based on the sequencing reads derived from the cancer DNA and DNA derived from healthy tissue of the cancer patient. The method of claim 1 or 6, wherein at least 99.99% of the sequencing reads map to the target locus.

Images