HBGSA: Hydrogen Bond Graph with Self-Attention for Drug-Target Binding Affinity Prediction

2026-04-24

AI Summary

This study addresses critical limitations in current drug–target binding affinity prediction methods, which often neglect spatial geometric constraints, hydrogen bonding characteristics, and the correlation between predicted scores and true affinities—hindering the identification of high-affinity compounds. To overcome these issues, this work proposes a novel approach that explicitly models the spatial topology of hydrogen bonds by constructing a hydrogen bond graph, integrates graph neural networks with self-attention mechanisms, and introduces a Pearson correlation loss to enhance prediction consistency. Evaluated on the PDBbind Core Set and CSAR-HiQ benchmarks, the method significantly outperforms existing baselines. Ablation studies confirm the effectiveness of both hydrogen bond modeling and the correlation-aware loss, demonstrating superior generalization and virtual screening performance.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Accurate prediction of drug-target binding affinity accelerates drug discovery by prioritizing compounds for experimental validation. Current methods face three limitations: sequence-based approaches discard spatial geometric constraints, structure-based methods fail to exploit hydrogen bond features, and conventional loss functions neglect prediction-target correlation, a key factor for identifying high-affinity compounds in virtual screening. We developed HBGSA (Hydrogen Bond Graph with Self-Attention), a 3.06M-parameter model that encodes hydrogen bond spatial features. HBGSA uses graph neural networks to model hydrogen bond spatial topology with self-attention enhancement and Pearson correlation loss. Experimental results on PDBbind Core Set and CSAR-HiQ dataset demonstrate that HBGSA outperforms baseline methods with strong generalization capability. Ablation studies confirm the effectiveness of hydrogen bond modeling and Pearson correlation loss.

Problem

• drug-target binding affinity
• hydrogen bond
• spatial geometry
• prediction-target correlation
• virtual screening

Innovation

• hydrogen bond graph
• self-attention
• binding affinity prediction
• graph neural network
• Pearson correlation loss