Surf2Spot: A Surface-Informed Geometry-Aware Model for Predicting Binder and Nanobody Design Hotspots
Zhao, A.; Li, Y.; Wu, T.; Fu, R.; Wang, Z.; Xu, Y.; Han, X.; Lou, Y.; Wang, W.; Yan, J.; Wang, X.
Show abstract
Protein-protein interactions (PPIs) and nanobody-antigen interactions (NAIs) play essential roles in cellular function, yet accurate hotspot prediction remains challenging. We present Surf2Spot, a deep learning framework that integrates sequence embeddings, structural features, and protein surface properties to predict interaction hotspots. By jointly modeling structural and physicochemical determinants, Surf2Spot achieves superior performance on curated PPI and NAI datasets, improving F1-scores by 45.9-46.9% and AUPRC by 43.2% over existing methods. Case studies on NbPDS1 and VdPDA1 demonstrate that Surf2Spot accurately recovers experimentally validated hotspots clustered within functional domains. When incorporated into binder design pipelines (RFdiffusion and BindCraft), Surf2Spot-guided designs yielded a 4-fold increase in successful design throughput and enhanced binding affinities compared to baseline strategies. These results establish Surf2Spot as a powerful tool for hotspot discovery and rational protein engineering.
Matching journals
The top 4 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- COLLAPSE: A representation learning framework for identification and characterization of protein structural sites 98%
- Evaluation of AlphaFold Antibody-Antigen Modeling with Implications for Improving Predictive Accuracy 97%
- Neural Network-Derived Potts Models for Structure-Based Protein Design using Backbone Atomic Coordinates and Tertiary Motifs 97%
Similar papers in this journal
- Predicting structures of large protein assemblies using combinatorial assembly algorithm and AlphaFold2 97%
- Direct prediction of intrinsically disordered protein conformational properties from sequence 97%
- Sliding Window INteraction Grammar (SWING): a generalized interaction language model for peptide and protein interactions 96%
Similar papers in this journal
- Deep Local Analysis deconstructs protein-protein interfaces and accurately estimates binding affinity changes upon mutation 96%
- Deep Local Analysis evaluates protein docking conformations with locally oriented cubes 96%
- ProBASS: a language model with sequence and structural features for predicting the effect of mutations on binding affinity 96%
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.