Surface-based Molecular Design with Multi-modal Flow Matching

2025-08-03

Knowledge Discovery and Data Mining

AI Summary

This work addresses a critical limitation in current therapeutic peptide design approaches, which often neglect the pivotal role of molecular surface properties in protein–protein interactions, thereby constraining the accuracy of binding prediction and peptide generation. To overcome this, we propose SurfFlow—the first multimodal conditional flow matching (CFM) generative model that explicitly integrates molecular surface geometry and biochemical features into de novo peptide design. SurfFlow jointly optimizes peptide sequence, structure, and surface characteristics to enable all-atom-level co-design of peptides and their receptors. Evaluated on the PepMerge benchmark, SurfFlow consistently outperforms existing all-atom generative models across all metrics, demonstrating the essential contribution of surface information to enhancing peptide binding affinity and specificity.

Technology Category

• Natural Language Processing: Generation
• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Therapeutic peptides show promise in targeting previously undruggable binding sites, with recent advancements in deep generative models enabling full-atom peptide co-design for specific protein receptors. However, the critical role of molecular surfaces in protein-protein interactions (PPIs) has been underexplored. To bridge this gap, we propose an omni-design peptides generation paradigm, called SurfFlow, a novel surface-based generative algorithm that enables comprehensive co-design of sequence, structure, and surface for peptides. SurfFlow employs a multi-modality conditional flow matching (CFM) architecture to learn distributions of surface geometries and biochemical properties, enhancing peptide binding accuracy. Evaluated on the comprehensive PepMerge benchmark, SurfFlow consistently outperforms full-atom baselines across all metrics. These results highlight the advantages of considering molecular surfaces in de novo peptide discovery and demonstrate the potential of integrating multiple protein modalities for more effective therapeutic peptide discovery.

Problem

• molecular surfaces
• protein-protein interactions
• therapeutic peptide design
• de novo peptide discovery
• binding accuracy

Innovation

• surface-based design
• multi-modal flow matching
• peptide co-design
• conditional flow matching
• molecular surface