This is the GitHub web for the paper "Drug Molecular Representations for Drug Response Predictions: A Comprehensive Investigation via Machine Learning Methods".
SRMF: MatLab R2023a, 64-bit;
ADRML: Python 3.7.14;
HiDRA: Python 3.9.7;
PathDSP: Python 3.9.7;
PaccMann: Python 3.9.7;
Source code: https://github.com/linwang1982/SRMF.
Source code: https://github.com/fahmadimoughari/AdrML.
Source code: https://github.com/TangYiChing/PathDSP.
The source code for HiDRA can be downloaded from https://pubs.acs.org/doi/10.1021/acs.jcim.1c00706?goto=supporting-info, in the supporting information section.
Source code: https://github.com/PaccMann/paccmann_predictor.
All preprocessing, training / predicting and analysis codes and data are available for each model under their folder named after them.
"Plots" is the folder for the code of plots shown in our paper.
256b, 512b, 1024b: refers to 256-bit, 512-bit and 1024-bit Morgan Fingerprints respectively. pbfp: refers PubChem fingerprints. mp(sometimes left as blank), mc, md: refers to mask pair/combination, mask cell, and mask drug splitting strategy respectively.