Fast protein backbone generation with SE(3) flow matching

2023/10/08 by Jason Yim, Andrew Campbell, Andrew M. Campbell +21 · 1 voice · 20 citations
Biochemistry, Genetics and Molecular Biology · Decision Sciences · Physics and Astronomy · #Model Reduction and Neural Networks #Protein Structure and Dynamics #Scientific Computing and Data Management #q-bio.QM

paper · pdf · doi:10.48550/arxiv.2310.05297

openalex publication_date 2023/10/08 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

We present FrameFlow, a method for fast protein backbone generation using SE(3) flow matching. Specifically, we adapt FrameDiff, a state-of-the-art diffusion model, to the flow-matching generative modeling paradigm. We show how flow matching can be applied on SE(3) and propose modifications during training to effectively learn the vector field. Compared to FrameDiff, FrameFlow requires five times fewer sampling timesteps while achieving two fold better designability. The ability to generate high quality protein samples at a fraction of the cost of previous methods paves the way towards more efficient generative models in de novo protein design.

Cited by

• Consistent Synthetic Sequences Unlock Structural Diversity in Fully Atomistic De Novo Protein Design
• ApexGen: Simultaneous design of peptide binder sequence and structure for target proteins
• EnzyControl: Adding Functional and Substrate-Specific Control for Enzyme Backbone Generation
• BioPipelines: Accessible Computational Protein and Ligand Design for Chemical Biologists
• ODesign: A World Model for Biomolecular Interaction Design
• Flows, straight but not so fast: Exploring the design space of Rectified Flows in Protein Design
• ProteinAE: Protein Diffusion Autoencoders for Structure Encoding
• Matching the Optimal Denoiser in Point Cloud Diffusion with (Improved) Rotational Alignment
• Distilled Protein Backbone Generation
• Riemannian Consistency Model
• Flow Autoencoders are Effective Protein Tokenizers
• Let Physics Guide Your Protein Flows: Topology-aware Unfolding and Generation
• Space Group Conditional Flow Matching
• High-Affinity Protein Binder Design via Flow Matching and In Silico Maturation
• Multi-domain Distribution Learning for De Novo Drug Design
• Monte Carlo Tree Diffusion with Multiple Experts for Protein Design
• Challenges in Non-Polymeric Crystal Structure Prediction: Why a Geometric, Permutation-Invariant Loss is Needed
• Energy-Based Flow Matching for Generating 3D Molecular Structure
• Sesame: Opening the door to protein pockets
• FlowMol3: Flow Matching for 3D De Novo Small-Molecule Generation

Discussions

• “Fast protein backbone generation with SE(3) flow matching” 🧶🧬 Shows that flow matching requires fewer timesteps than diffusion when generating novel protein backbones arxiv.org/abs/2310.05297 [bsky, 5 points, 0 comments]

Related

• SE(3)-Stochastic Flow Matching for Protein Backbone Generation
• Flows, straight but not so fast: Exploring the design space of Rectified Flows in Protein Design
• Sequence-Augmented SE(3)-Flow Matching For Conditional Protein Backbone Generation
• SE(3) diffusion model with application to protein backbone generation
• Protein-SE(3): Benchmarking SE(3)-based Generative Models for Protein Structure Design