Probabilistic macrochemical modeling is a general methodology that builds on macrochemical characterizations of microbial growth (Heijen and Van Dijken, 1992). With it, we can build probabilistic models that infer from all available experimental data globally consistent and accurate estimates of microbial quantities of interest (e.g., biomass yields).
Specifically, this repository contains the code used to generate the validation data and perform the analysis in the article,
- Antonio R. Paiva and Giovanni Pilloni (2022). Inferring Microbial Biomass Yield and Cell Weight using Probabilistic Macrochemical Modeling. IEEE/ACM Transactions on Computational Biology and Bioinformatics
DOI: 10.1109/TCBB.2021.3139290 | arXiv preprint arXiv:2010.02759
The code has been tested in Python 3.7 and PyStan 2.19. Note that, while newer versions of Python 3 should work, PyStan 3 is not backwards compatible and would require a few changes to the code.
The repository comprises 4 main files:
-
srbsim.pyimplements the class used to generate the simulated data used in the paper. This is called in each of the Jupyter notebooks. -
chem_model1_constCellWeight.ipynbimplements all analysis pertaining to Scenario 1 described in the paper. -
chem_model2_changingCellWeights.ipynbimplements all analysis pertaining to Scenario 2 described in the paper. -
chem_model3_changingCellWeights_biomassSensitivity.ipynbtests, with respect to Scenario 2, the sensitivity of the results to the generic biomass composition formula used in the paper. This analysis is described in the paper supplementary materials, Section S1.
This code is being shared under an MIT license.