PPFlow: Target-aware Peptide Design with Torsional Flow Matching
This work addresses the need for improved peptide drug design methods for pharmaceutical applications, representing a novel approach in this domain.
The authors tackled the problem of AI-assisted peptide drug discovery by proposing PPFlow, a target-aware peptide design method based on conditional flow matching on torus manifolds, which achieved state-of-the-art performance in peptide generation and optimization tasks.
Therapeutic peptides have proven to have great pharmaceutical value and potential in recent decades. However, methods of AI-assisted peptide drug discovery are not fully explored. To fill the gap, we propose a target-aware peptide design method called \textsc{PPFlow}, based on conditional flow matching on torus manifolds, to model the internal geometries of torsion angles for the peptide structure design. Besides, we establish a protein-peptide binding dataset named PPBench2024 to fill the void of massive data for the task of structure-based peptide drug design and to allow the training of deep learning methods. Extensive experiments show that PPFlow reaches state-of-the-art performance in tasks of peptide drug generation and optimization in comparison with baseline models, and can be generalized to other tasks including docking and side-chain packing.