Thanks for the feedback. GFF/GTF/GFF3 support is pretty simplistic, all handled by the same parser, and relatively untested. On my GFF3 test files the parser picks up gene names correctly, but I'll need to modify it to handle the GTF syntax for specifying gene names and ensure the names are carried over to target labels.

The problem is in this line:

Judging from the links in the Specs section in gff.py, there seems to be no universal agreed upon standard for specifying gene names. For ensmbl it should be tag = 'gene_name='.

Maybe a solution would be to have Name as default but then allow the user to optionally specify a different tag name?

I have come up with a possible solution in this branch. Not sure if it fits into the general idea of how your code is structured, but it works (for GFF files):

$ cnvkit.py target --annotate gencode.v27.annotation.gff3 test.bed
Detected file format: bed
Applying annotations as target names
Detected file format: gff
chr17	43124026	43124115	-

$ cnvkit.py target --annotate gencode.v27.annotation.gff3 --tag gene_name test.bed
Detected file format: bed
Applying annotations as target names
Detected file format: gff
chr17	43124026	43124115	BRCA1

Thanks for the example code. Rather than add the --tag option to each command that can accept input regions in BED/GFF/etc. format, I think a safer workflow is to use the skg_convert.py script to convert GFF2/GTF/GFF3 to BED, adding a --gff-tag option there to allow looking at nonstandard tags, or specifying one tag if multiple "standard" tags are present in the file. Then use the resulting BED file with target and other commands.

The command line for that script is:

skg_convert.py -f gff -t bed4 gencode.v27.annotation.gff3 -o gencode.v27.bed
cnvkit.py target --annotate gencode.v27.bed test.bed

With the latest commits to make GFF parsing more permissive, this should work on the gencode files. (It takes about a minute on my machine.) I have not added the --gff-tag option to skg_convert.py yet.

I've added some features to skg_convert.py to help extract the right info from your GFF file. Use it like:

skg_convert.py gencode.v27.annotation.gff3 -f gff -t bed4 -o gencode.v27.bed --gff-type exon --gff-tag gene_name --merge

(Use -h for command line option descriptions, as always.)