Analysis code, processed data, and notebooks accompanying the BIRDMAn (Bayesian Inferential Regression for Differential Microbiome Analysis) manuscript.
BIRDMAn is a framework for fitting Bayesian differential abundance models to microbiome count data, built on top of CmdStanPy. This repository contains the end-to-end analyses used to evaluate BIRDMAn in the manuscript, including the scripts used to fit models on a SLURM cluster, the processed input data, and the Jupyter notebooks that generate the manuscript figures. BIRDMAn itself is available at biocore/birdman.
Runtime dependencies include birdman, cmdstanpy, biom-format, arviz, pandas, numpy, and matplotlib. We suggest installing cmdstanpy from conda-forge.
If you use this code or the processed data, please cite the BIRDMAn manuscript:
- Gibraan Rahman — original author (
gibsramen@gmail.com) - Lucas Patel — co-maintainer (
lpatel@ucsd.edu) - Yang Chen — co-maintainer (
yac027@ucsd.edu)
- Code (
src/,scripts/,notebooks/): BSD 3-Clause — seeLICENSE. - Processed data (
data/): CC0 1.0 Universal Public Domain Dedication — seedata/LICENSE.