consistency-paper

Code and documentation for the consitency paper

Installation

Prerequisites:

• Python 3.5 or 3.6
• Python package nwtools. You can install it with: pip install -e https://github.com/research-Dafne/network-tools.

scripts

run_consistency.py

Read a node file and edge file, and run consensus clustering for a specific threshold, and calculate edge and node consistency scores.

Example usage: python scripts/run_consistency.py data/edges.csv data/nodes.csv

Usage: run_consistency.py [OPTIONS] EDGE_FILE NODE_FILE

Options:
  -sn, --sepn TEXT  seperator for node file
  -se, --sepe TEXT  seperator for edges file
  -i, --idcol TEXT  name of ID column in nodes header
  -w, --weightcol TEXT  name of weight column in edges header
  -t, --threshold FLOAT threshold for consensus clustering
  -n, --nriter INTEGER  number of initial clusterings
  -o, --out_dir PATH
  --help

consensus_clustering.py

Runs consensus clustering for different thresholds, with the same initial partitions that it reads from a file on DATAPATH

Usage: consensus_clustering.py [OPTIONS] DATAPATH

Options:
  -s, --stepsize FLOAT  stepsize for threshold list
  -o, --out_dir PATH
  --help

Notebooks

There is a notebook with the plots in the paper, as well as some further exploration, on an example data set.