On Machine Learning Approaches for Protein-Ligand Binding Affinity Prediction

TL;DR

This paper benchmarks various machine learning models, including 2D and 3D neural networks, for predicting protein-ligand binding affinity, revealing that simpler models can outperform complex ones in certain scenarios and that combining approaches enhances active learning.

Contribution

It provides a comprehensive comparison of classical and advanced ML models for binding affinity prediction, highlighting the benefits of pretraining and multi-modal approaches in drug discovery.

Findings

Simpler models can outperform complex models in specific tasks.

Pre-trained 3D models excel in data-scarce scenarios with structural info.

Combining 2D and 3D models improves active learning performance.

Abstract

Binding affinity optimization is crucial in early-stage drug discovery. While numerous machine learning methods exist for predicting ligand potency, their comparative efficacy remains unclear. This study evaluates the performance of classical tree-based models and advanced neural networks in protein-ligand binding affinity prediction. Our comprehensive benchmarking encompasses 2D models utilizing ligand-only RDKit embeddings and Large Language Model (LLM) ligand representations, as well as 3D neural networks incorporating bound protein-ligand conformations. We assess these models across multiple standard datasets, examining various predictive scenarios including classification, ranking, regression, and active learning. Results indicate that simpler models can surpass more complex ones in specific tasks, while 3D models leveraging structural information become increasingly competitive with larger training datasets containing compounds with labelled affinity data against multiple targets. Pre-trained 3D models, by incorporating protein pocket environments, demonstrate significant advantages in data-scarce scenarios for specific binding pockets. Additionally, LLM pretraining on 2D ligand data enhances complex model performance, providing versatile embeddings that outperform traditional RDKit features in computational efficiency. Finally, we show that combining 2D and 3D model strengths improves active learning outcomes beyond current state-of-the-art approaches. These findings offer valuable insights for optimizing machine learning strategies in drug discovery pipelines.