OnionNet-2: A Convolutional Neural Network Model for Predicting Protein-Ligand Binding Affinity based on Residue-Atom Contacting Shells

TL;DR

OnionNet-2 is a convolutional neural network model that predicts protein-ligand binding affinity using residue-atom contact shells, outperforming previous models on benchmark datasets and verified with diverse data sources.

Contribution

This paper introduces OnionNet-2, a novel CNN-based scoring function that effectively predicts binding free energy using contact shells, improving accuracy and efficiency over existing models.

Findings

Outperforms existing models on CASF-2016 and CASF-2013 datasets.

Successfully predicts binding affinity on non-experimental docking decoys.

Demonstrates robustness across multiple high-quality datasets.

Abstract

One key task in virtual screening is to accurately predict the binding affinity ($\triangle$$G$) of protein-ligand complexes. Recently, deep learning (DL) has significantly increased the predicting accuracy of scoring functions due to the extraordinary ability of DL to extract useful features from raw data. Nevertheless, more efforts still need to be paid in many aspects, for the aim of increasing prediction accuracy and decreasing computational cost. In this study, we proposed a simple scoring function (called OnionNet-2) based on convolutional neural network to predict $\triangle$$G$. The protein-ligand interactions are characterized by the number of contacts between protein residues and ligand atoms in multiple distance shells. Compared to published models, the efficacy of OnionNet-2 is demonstrated to be the best for two widely used datasets CASF-2016 and CASF-2013 benchmarks. The OnionNet-2 model was further verified by non-experimental decoy structures from docking program and the CSAR NRC-HiQ data set (a high-quality data set provided by CSAR), which showed great success. Thus, our study provides a simple but efficient scoring function for predicting protein-ligand binding free energy.