SurfDesign: Effective Protein Design on Molecular Surfaces

2026/05/25 by Fang Wu, Shuting Jin, Xiangru Tang +5 · 1 voice · 2 citations
Biochemistry, Genetics and Molecular Biology · Computer Science · #q-bio.BM #cs.AI #cs.CE

paper · pdf

Abstract

Protein function is largely determined by molecular surface geometry and physicochemical complementarity, yet most protein design methods condition only on backbone structure. We introduce SurfDesign, a surface-conditioned protein design framework that models molecular surfaces as continuous geometric manifolds and integrates them with pretrained protein language models. SurfDesign employs surface-based equivariant message passing to capture surface normals, curvature, and directional geometry, together with a parameter-efficient fine-tuning strategy. Focusing on functional protein design, we show that SurfDesign consistently outperforms prior surface-conditioned and backbone-only methods on de novo binder and enzyme design benchmarks. We also report strong performance on inverse-folding benchmarks as a diagnostic of structural compatibility. Our results highlight manifold-aware surface representations as a principled foundation for functional protein and enzyme design. Code is available at https://github.com/smiles724/SurfDesign.

Citations

• PRISM: Enhancing Protein Inverse Folding through Fine-Grained Retrieval on Structure-Sequence Multimodal Representations
• DeepSearch: Overcome the Bottleneck of Reinforcement Learning with Verifiable Rewards via Monte Carlo Tree Search
• Protein Inverse Folding From Structure Feedback
• Mask prior-guided denoising diffusion improves inverse protein folding
• Reinforcement learning on structure-conditioned categorical diffusion for protein inverse folding
• UniIF: Unified Molecule Inverse Folding
• Accurate structure prediction of biomolecular interactions with AlphaFold 3
• SurfPro: Functional Protein Design Based on Continuous Surface
• Simple, Efficient and Scalable Structure-aware Adapter Boosts Protein Language Models
• InstructPLM: Aligning Protein Language Models to Follow Protein Structure Instructions
• Diffusion Language Models Are Versatile Protein Learners
• PDB-Struct: A Comprehensive Benchmark for Structure-based Protein Design
• A Hierarchical Training Paradigm for Antibody Structure-sequence Co-design
• Graph Denoising Diffusion for Inverse Protein Folding
• GeoMAE: Masked Geometric Target Prediction for Self-supervised Point Cloud Pre-Training
• EquiPocket: an E(3)-Equivariant Geometric Graph Neural Network for Ligand Binding Site Prediction
• SE(3) diffusion model with application to protein backbone generation
• Structure-informed Language Models Are Protein Designers
• Integration of Pre-trained Protein Language Models into Geometric Deep Learning Networks
• PiFold: Toward effective and efficient protein inverse folding
• Learning inverse folding from millions of predicted structures
• ProGen2: Exploring the Boundaries of Protein Language Models
• Exploring evolution-aware & -free protein language models as protein function predictors
• Diffusion probabilistic modeling of protein backbones in 3D for the motif-scaffolding problem
• Robust deep learning–based protein sequence design using ProteinMPNN
• Generative De Novo Protein Design with Global Context
• DiffMD: A Geometric Diffusion Model for Molecular Dynamics Simulations
• Multi-Scale Representation Learning on Proteins
• Protein Representation Learning by Geometric Structure Pretraining
• Improving Molecular Representation Learning with Metric Learning-enhanced Optimal Transport
• Fast and accurate protein structure search with Foldseek
• AlphaDesign: A graph protein design method and benchmark on AlphaFoldDB
• OntoProtein: Protein Pretraining With Gene Ontology Embedding
• Step-unrolled Denoising Autoencoders for Text Generation
• КОМПЬЮТЕРНАЯ ТОМОГРАФИЯ – ОСНОВА ОПТИМИЗАЦИИ МИНИТОРАКОТОМНОГО ДОСТУПА
• Molformer: Motif-based Transformer on 3D Heterogeneous Molecular Graphs
• Highly accurate protein structure prediction for the human proteome
• Highly accurate protein structure prediction with AlphaFold
• LoRA: Low-Rank Adaptation of Large Language Models
• The Power of Scale for Parameter-Efficient Prompt Tuning
• E(n) Equivariant Graph Neural Networks
• Fast and Uncertainty-Aware Directional Message Passing for Non-Equilibrium Molecules
• Score-Based Generative Modeling through Stochastic Differential Equations
• Learning from Protein Structure with Geometric Vector Perceptrons
• ProtTrans: Towards Cracking the Language of Life's Code Through Self-Supervised Deep Learning and High Performance Computing
• ProtTrans: Toward Understanding the Language of Life Through Self-Supervised Learning
• Directional Message Passing for Molecular Graphs
• Scaling Laws for Neural Language Models
• Evaluating Protein Transfer Learning with TAPE
• DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation
• Activation Functions: Comparison of trends in Practice and Research for Deep Learning
• Computational Protein Design with Deep Learning Neural Networks
• The limits of protein sequence comparison?
• Scoring function for automated assessment of protein structure template quality
• Dictionary of protein secondary structure: Pattern recognition of hydrogen‐bonded and geometrical features
• Discovering the Representation Bottleneck of Graph Neural Networks

Discussions

• Fang Wu, Shuting Jin, Xiangru Tang, Mark Gerstein, Xiangxiang Zeng, Yejin Choi, Jure Leskovec, Jinbo Xu: SurfDesign: Effective Protein Design on Molecular Surfaces https://arxiv.org/abs/2606.07567 htt [bsky, 0 points, 0 comments]

Related

• SurfPro: Functional Protein Design Based on Continuous Surface
• Joint Design of Protein Surface and Structure Using a Diffusion Bridge Model
• De novoprotein design, a retrospective
• ProteinWeaver: A Divide-and-Assembly Approach for Protein Backbone Design
• DS-ProGen: A Dual-Structure Deep Language Model for Functional Protein Design