Application of evolutionary coupling inference to study of bacterial epistasis, as described in 'Interacting networks of resistance, virulence and core machinery genes identified by genome-wide epistasis analysis' by Marcin J Skwark, Nicholas J Croucher, Santeri Puranen, Claire Chewapreecha, Maiju Pesonen, Ying Ying Xu, Paul Turner, Simon R Harris, Stephen B. Beres, James M. Musser, Julian Parkhill, Stephen D Bentley, Erik Aurell, Jukka Corander.
This repository is a snapshot of software used to obtain results presented in the paper.
In order to successfully conduct this analysis, one needs:
- Python interpreter
- MATLAB installation
- an alignment of whole genomes for analysis
- (optionally) BLAST for analysis of the results on protein level
The computational procedure comprises the following steps:
- Analysis of the allelle distribution in the input genomes, filtering and encoding information for subsequent steps (
00datafiltering) - Generation of pseudo-alignments for coupling inference (
01sampling) - Coupling inference (
02inference). Warning: this step is computationally intensive, but is amenable to parallel processing. - Extraction of information from inferred couplings (
03postprocessing). - (optional) Mapping the inferred couplings onto protein sequences (
04analysis)
It is possible to replace MATLAB with its open-source counterpart, Octave. We do not support this procedure and do not guarantee reproducibility of results.
- All changes necessary are to be done in 02inference/genomeDCA/plmDCA_asymmetric.m
- Remove reference to
matlabpool - Compile
*.cfiles withmkoctfile --mex - Ensure that resultant file are in Octave's search path