Learning Protein-Ligand Binding in Hyperbolic Space

TL;DR

HypSeek introduces a hyperbolic space embedding framework for protein-ligand interactions, capturing hierarchical and subtle affinity variations more effectively than Euclidean models, improving virtual screening and affinity ranking.

Contribution

This work presents HypSeek, the first hyperbolic embedding method for protein-ligand binding prediction, unifying virtual screening and affinity ranking with enhanced modeling of complex molecular relationships.

Findings

Improves virtual screening early enrichment from 42.63 to 51.44.

Enhances affinity ranking correlation from 0.5774 to 0.7239.

Effectively models activity cliffs and subtle affinity differences.

Abstract

Protein-ligand binding prediction is central to virtual screening and affinity ranking, two fundamental tasks in drug discovery. While recent retrieval-based methods embed ligands and protein pockets into Euclidean space for similarity-based search, the geometry of Euclidean embeddings often fails to capture the hierarchical structure and fine-grained affinity variations intrinsic to molecular interactions. In this work, we propose HypSeek, a hyperbolic representation learning framework that embeds ligands, protein pockets, and sequences into Lorentz-model hyperbolic space. By leveraging the exponential geometry and negative curvature of hyperbolic space, HypSeek enables expressive, affinity-sensitive embeddings that can effectively model both global activity and subtle functional differences-particularly in challenging cases such as activity cliffs, where structurally similar ligands exhibit large affinity gaps. Our mode unifies virtual screening and affinity ranking in a single framework, introducing a protein-guided three-tower architecture to enhance representational structure. HypSeek improves early enrichment in virtual screening on DUD-E from 42.63 to 51.44 (+20.7%) and affinity ranking correlation on JACS from 0.5774 to 0.7239 (+25.4%), demonstrating the benefits of hyperbolic geometry across both tasks and highlighting its potential as a powerful inductive bias for protein-ligand modeling.