Surface-based Peptide Design with Multi-modal Flow Matching
Lists
Only you see this
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 \emph{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 \emph{de novo} peptide discovery and demonstrate the potential of integrating multiple protein modalities for more effective therapeutic peptide discovery. Anonymous codes are available at~\url{https://anonymous.4open.science/r/SurfFlow-880B/}.
BibTeX
@misc{wu2025surface,
author = {Fang Wu and Shuting Jin and xiangxiang Zeng and Jure Leskovec and Jinbo Xu},
title = {{Surface-based Peptide Design with Multi-modal Flow Matching}},
year = {2025},
url = {https://openreview.net/forum?id=MeCPwqrm19},
howpublished = {OpenReview},
note = {Submitted to ICLR 2025},
}