Data-centric training enables meaningful interaction learning in protein–ligand binding affinity prediction
Preprint
- 6 October 2025
- preprint
- Published by American Chemical Society (ACS) in ChemRxiv
Abstract
Predicting protein–ligand binding affinity remains a fundamental challenge in drug discovery, where machine learning models often struggle with rotational sensitivity, overfitting, and poor generalization. Here, we present a convolutional neural network (CNN)–based scoring function that emphasizes the careful design of the training strategy over architectural complexity. Systematic evaluation of data augmentation strategies shows that random rotation, unlike commonly used 90-degree rotations, effectively mitigates rotational variance. We also introduce a simple regularization technique that reduces reliance on isolated protein or ligand features and, together with random rotation, promotes learning of meaningful intermolecular interaction patterns. The resulting model, built on a simple architecture, generalizes across diverse test conditions, including structurally disjoint splits, and matches or exceeds the performance of state-of-the-art methods. Through systematic ablation studies and interpretability experiments, we provide insights into model behaviour, highlighting the importance of data-centric design in developing reliable scoring functions.Keywords
All Related Versions
This publication has 56 references indexed in Scilit:
- Scaffold Splits Overestimate Virtual Screening Performance2024
- Enhancing protein‐ligand binding affinity prediction through sequential fusion of graph and convolutional neural networksJournal of Computational Chemistry, 2024
- Modern machine‐learning for binding affinity estimation of protein–ligand complexes: Progress, opportunities, and challengesWIREs Computational Molecular Science, 2024
- DockThor-VS: A Free Platform for Receptor-Ligand Virtual ScreeningJournal of Molecular Biology, 2024
- Docking-based generative approaches in the search for new drug candidatesDrug Discovery Today, 2022
- Complex machine learning model needs complex testing: Examining predictability of molecular binding affinity by a graph neural networkJournal of Computational Chemistry, 2022
- Data Set Augmentation Allows Deep Learning-Based Virtual Screening to Better Generalize to Unseen Target Classes and Highlight Important Binding InteractionsJournal of Chemical Information and Modeling, 2020
- Forging the Basis for Developing Protein–Ligand Interaction Scoring FunctionsAccounts of Chemical Research, 2017
- Can we trust docking results? Evaluation of seven commonly used programs on PDBbind databaseJournal of Computational Chemistry, 2010
- Epik: a software program for pK a prediction and protonation state generation for drug-like moleculesJournal of Computer-Aided Molecular Design, 2007