This repository contains code and data for simulating pairwise microbial interactions in the human gut microbiome. The paper is available at:
The code and details for SIMBA's GNN architecture are available at: [https://github.com/msparsa/simba/]
- Input: Microbial abundance data from a high-fiber diet study (Diener et al., 2020)
- Models: GEMs from AGORA v1.03 (Heinken et al., 2020)
- Simulations: 2,850 pairwise co-cultures simulated using COBRApy under anaerobic, high-fiber diet
- Flux Sampling:
optGpSampler - Pathway Analysis: Functional fingerprints derived from KEGG and VMH annotations
- Multi-compartment models with a shared lumen for metabolite exchange
- Diet-constrained flux simulations reflecting gut conditions
- Quantification of cross-fed metabolites
- Functional analysis of microbe-microbe interactions
- Pathway activity scoring based on KEGG and VMH annotations
- Python 3.8+
- COBRApy
- 'optGpSampler'
- numpy, pandas, scipy, matplotlib, seaborn
Run the simulation pipeline:
cd sampling
python main_sampling.pyEnsure all required model files and input data (e.g., GEMs, diet constraints) are placed in the correct subdirectories.
This pipeline complements the SIMBA framework, which models microbial community behavior using graph neural networks and metabolic modeling.
If you use this repository, please cite the associated publication:
This project is licensed under the Creative Commons Attribution 4.0 International License.