AlphaFold Meets Flow Matching for Generating Protein Ensembles

TL;DR

This paper introduces AlphaFlow and ESMFlow, flow-based generative models fine-tuned from AlphaFold and ESMFold to efficiently generate diverse protein conformational ensembles, capturing flexibility and dynamics more effectively than traditional methods.

Contribution

It develops a novel flow matching framework applied to protein structure prediction models, enabling fast and accurate sampling of protein conformational landscapes.

Findings

Outperforms AlphaFold with MSA subsampling in precision and diversity.

Accurately captures conformational flexibility and ensemble observables.

Provides a faster alternative to physics-based MD simulations.

Abstract

The biological functions of proteins often depend on dynamic structural ensembles. In this work, we develop a flow-based generative modeling approach for learning and sampling the conformational landscapes of proteins. We repurpose highly accurate single-state predictors such as AlphaFold and ESMFold and fine-tune them under a custom flow matching framework to obtain sequence-conditoned generative models of protein structure called AlphaFlow and ESMFlow. When trained and evaluated on the PDB, our method provides a superior combination of precision and diversity compared to AlphaFold with MSA subsampling. When further trained on ensembles from all-atom MD, our method accurately captures conformational flexibility, positional distributions, and higher-order ensemble observables for unseen proteins. Moreover, our method can diversify a static PDB structure with faster wall-clock convergence to certain equilibrium properties than replicate MD trajectories, demonstrating its potential as a proxy for expensive physics-based simulations. Code is available at https://github.com/bjing2016/alphaflow.