Some python tools for making GO term prediction using protein networks
Start by creating the python virtual environment by running in the project:
% python3 -m venv .venv
% source .venv/bin/activate
% pip install -r requirements.txt
To fetch the STRING data, run
% python3 fetch_data.py
Then to process the STRING data to get DSD output, adjacency matrix, and GO terms, run
% python3 networkgraph.py {LINK_FILE} {GO_FILE}
Where the link and go entry files are both in the data folder.
This program results in three csv files in the results folder.
To run the ML model on these csv files, run
% python3 NetworksML.py --adj-path {ADJACENCY_MATRIX_FILE} --go-path {GO_MATRIX_FILE}
This creates descriptive statistics plots, accuracy and ROC curve plots, and a summary report in the results folder.