AlphaFold Meets Flow Matching for Generating Protein Ensembles

Bowen Jing, Bonnie Berger, Tommi Jaakkola
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:22277-22303, 2024.

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-conditioned 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.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-jing24a, title = {{A}lpha{F}old Meets Flow Matching for Generating Protein Ensembles}, author = {Jing, Bowen and Berger, Bonnie and Jaakkola, Tommi}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {22277--22303}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/jing24a/jing24a.pdf}, url = {https://proceedings.mlr.press/v235/jing24a.html}, 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-conditioned 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.} }
Endnote
%0 Conference Paper %T AlphaFold Meets Flow Matching for Generating Protein Ensembles %A Bowen Jing %A Bonnie Berger %A Tommi Jaakkola %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-jing24a %I PMLR %P 22277--22303 %U https://proceedings.mlr.press/v235/jing24a.html %V 235 %X 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-conditioned 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.
APA
Jing, B., Berger, B. & Jaakkola, T.. (2024). AlphaFold Meets Flow Matching for Generating Protein Ensembles. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:22277-22303 Available from https://proceedings.mlr.press/v235/jing24a.html.

Related Material