Geometric Pocket-Centric Protein Encoding for Polypharmacology-Guided Multi-Target Drug Design

Polypharmacology provides a powerful strategy for treating complex diseases, but identifying molecules that simultaneously satisfy coupled constraints across multiple biological targets remains difficult. Existing methods typically model protein pockets in isolation and struggle to jointly account for multiple heterogeneous binding sites when designing a single shared ligand. To address these limitations, we propose a pocket-structure-centric generative framework for polypharmacology. This framework introduces a novel protein topological representation that selectively masks ligand-irrelevant residues while explicitly modeling backbone folding geometry and inter-residue spatial proximity within binding pockets. In addition, structural representations are jointly fused with amino acid and nucleotide sequences to capture their complementary information across targets. Experiments on COVID-19, schizophrenia, and tumor targets show that this framework generates valid candidates with significantly improved binding affinities compared to state-of-the-art methods.

Polypharmacology aims to treat complex diseases using a single drug that acts on multiple biological targets at the same time. However, designing such molecules remains difficult because different proteins often contain diverse binding pockets with competing structural requirements. We develop an Artificial Intelligence framework for polypharmacology guided molecular design that directly uses information from protein binding pockets during molecule generation. Rather than modeling targets independently, our method learns from multiple targets jointly and combines protein structural information with amino acid and nucleotide sequences. The framework emphasizes protein regions involved in molecular interactions while reducing the influence of unrelated structural regions. Experiments on targets associated with COVID-19, schizophrenia, and cancer show that our method generates valid candidate molecules with stronger binding affinities than existing approaches. These findings suggest that incorporating detailed protein information into generative models can improve polypharmacology driven drug discovery for complex diseases.