Structure-Aware Compound-Protein Affinity Prediction via Graph Neural Network with Group Lasso Regularization

TL;DR

This paper introduces a graph neural network framework with group lasso regularization for predicting compound-protein affinity, especially focusing on activity cliffs, improving both accuracy and interpretability in drug discovery.

Contribution

The study presents a novel GNN-based approach with structure-aware loss functions and regularizations to enhance prediction accuracy and explainability for activity cliff data in drug discovery.

Findings

Improved affinity prediction accuracy with reduced RMSE and higher PCC.

Enhanced feature attribution and interpretability of molecular substructures.

Effective identification of critical molecular substructures influencing activity.

Abstract

Explainable artificial intelligence (XAI) approaches have been increasingly applied in drug discovery to learn molecular representations and identify substructures driving property predictions. However, building end-to-end explainable models for structure-activity relationship (SAR) modeling for compound property prediction faces many challenges, such as the limited number of compound-protein interaction activity data for specific protein targets, and plenty of subtle changes in molecular configuration sites significantly affecting molecular properties. We exploit pairs of molecules with activity cliffs that share scaffolds but differ at substituent sites, characterized by large potency differences for specific protein targets. We propose a framework by implementing graph neural networks (GNNs) to leverage property and structure information from activity cliff pairs to predict compound-protein affinity (i.e., half maximal inhibitory concentration, IC50). To enhance model performance and explainability, we train GNNs with structure-aware loss functions using group lasso and sparse group lasso regularizations, which prune and highlight molecular subgraphs relevant to activity differences. We applied this framework to activity cliff data of molecules targeting three proto-oncogene tyrosine-protein kinase Src proteins (PDB IDs: 1O42, 2H8H, 4MXO). Our approach improved property prediction by integrating common and uncommon node information with sparse group lasso, as reflected in reduced root mean squared error (RMSE) and improved Pearson's correlation coefficient (PCC). Applying regularizations also enhances feature attribution for GNN by boosting graph-level global direction scores and improving atom-level coloring accuracy. These advances strengthen model interpretability in drug discovery pipelines, particularly for identifying critical molecular substructures in lead optimization.