Code for Hall-Swan et al. 2017, "Detangling PPI networks to uncover functionally meaningful clusters".
Many scripts in this repository require numpy, pandas, and any recent version of scikit-learn.
Running our implementations of spectral clustering and Louvain clustering require the Python igraph package: pip install python-igraph or conda install -c conda-forge python-igraph.
All of the scripts should be run from the repo root.
To run DSD on the yeast network, and generate the filtered DSD matrices: ./scripts/dsd.sh
To run (spectral|walktrap|Louvain) clustering: ./scripts/clusters_(spectral|wt|louvain).sh
To calculate GO term enrichment (using the method described in the paper) for the generated clusters: ./scripts/calc_enrichment.sh
To generate tables and figures from enrichment results: ./scripts/tables_and_figures.sh
This repository is mostly for purposes of reproducing the experiments and results in the above paper. I'll try to respond to any issues posted on the GitHub issue tracker, but I won't be actively maintaining or adding functionality to the repo.