This repository includes the codes used in the work of Forman Persistent Ricci Curvature (FPRC) in Protein-Ligand Binding Affinity Predictions. This work has shown that FPRC delivers meaningful and interesting geometrical features in biomolecular data which is able to predict binding affinities with comparable performance to many existing traditional molecular descriptors and advanced mathematical based machine learning models. FPRC is able to provide discretized geometric molecular representation on simplicial complex level and generates intrinsic geometrical features in predicting binding affinities.
If you use this code in your research, please considering cite our paper:
@article{wee2021forman,
title={Forman persistent Ricci curvature (FPRC)-based machine learning models for protein--ligand binding affinity prediction},
author={Wee, JunJie and Xia, Kelin},
journal={Briefings in Bioinformatics},
year={2021}
}