Accurate structure prediction of biomolecular interactions with AlphaFold3
Josh Abramson ^ 17 Core Contributor, <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Jonas Adler ^ 17 Core Contributor, <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Jack Dunger ^ 1,7 Core Contributor, <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Richard Evans ^ 1,7 Core Contributor, <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Tim Green ^ 1,7 Core Contributor, <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Alexander Pritzel ^ 17 Core Contributor, <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Olaf Ronneberger ^ 17 Core Contributor, <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Lindsay Willmore ^ 1,7 Core Contributor, <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Andrew J Ballard ^ 1 Core Contributor, <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Joshua Bambrick ^ 2 Core Contributor, <a href="/isomorphic_labs">Isomorphic Labs</a>, <a href="/london">London</a>, UK.
Sebastian W Bodenstein <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
David A Evans <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Chia-Chun Hung ^ 1 Core Contributor, <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Michael O'Neill <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
David Reiman ^ 1 Core Contributor, <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Kathryn Tunyasuvunakool <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Zachary Wu <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Akvilè Žemgulytè <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Eirini Arvaniti ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Charles Beattie ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Ottavia Bertolli ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Alex Bridgland ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Alexey Cherepanov ^ 4 <a href="/isomorphic_labs">Isomorphic Labs</a>, <a href="/london">London</a>, UK.
Miles Congreve ^ 4 <a href="/isomorphic_labs">Isomorphic Labs</a>, <a href="/london">London</a>, UK.
Alexander I Cowen-Rivers ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Andrew Cowie ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Michael Figurnov ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Fabian B Fuchs ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Hannah Gladman ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Rishub Jain ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Yousuf A Khan ^ 3,5 Department of Molecular
Caroline M R Low ^ 4 <a href="/isomorphic_labs">Isomorphic Labs</a>, <a href="/london">London</a>, UK.
Kuba Perlin ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Anna Potapenko ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Pascal Savy ^ 4 <a href="/isomorphic_labs">Isomorphic Labs</a>, <a href="/london">London</a>, UK.
Sukhdeep Singh ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Adrian Stecula ^ 4 <a href="/isomorphic_labs">Isomorphic Labs</a>, <a href="/london">London</a>, UK.
Ashok Thillaisundaram ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Catherine Tong ^ 4 <a href="/isomorphic_labs">Isomorphic Labs</a>, <a href="/london">London</a>, UK.
Sergei Yakneen ^ 4 <a href="/isomorphic_labs">Isomorphic Labs</a>, <a href="/london">London</a>, UK.
Ellen D Zhong ^ 3,6 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Michal Zielinski ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Augustin Židek ^ 3 <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Victor Bapst ^ 1,8 Core Contributor, <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
Pushmeet Kohli ^ 1,8 Core Contributor, Google Deep Proposed Instructionsson, <a href="/london">London</a>, UK.
Max Jaderberg ^ 2,8 Core Contributor, <a href="/isomorphic_labs">Isomorphic Labs</a>, <a href="/london">London</a>, UK.
Demis Hassabis ^ 1,2,8 Core Contributor, <a href="/google_deepmind">Google DeepMind</a> and <a href="/isomorphic_labs">Isomorphic Labs</a>, <a href="/london">London</a>, UK.
John M Jumper ^ 1,8 Core Contributor, <a href="/google_deepmind">Google DeepMind</a>, <a href="/london">London</a>, UK.
1 Core Contributor, Google DeepMind, London, UK.
2 Core Contributor, Isomorphic Labs, London, UK.
3 Google DeepMind, London, UK.
4 Isomorphic Labs, London, UK.
5 Department of Molecular and Cellular Physiology, Stanford University, Stanford , CA, USA.
6 Department of Computer Science, Princeton University, Princeton, NJ, USA.
7 These authors contributed equally: Josh Abramson, Jonas Adler, Jack Dunger, Richard Evans, Tim Green, Alexander Pritzel, Olaf Ronneberger, Lindsay Willmore.
8 These authors jointly supervised this work: Victor Bapst, Pushmeet Kohli, Max Jaderberg, Demis Hassabis, John M. Jumper.
✉e-mail: jaderberg@isomorphiclabs.com; dhcontact@google.com; jumper@google.com
The introduction of AlphaFold $2^{1}$ has spurred a revolution in modelling the structure of proteins and their interactions, enabling a huge range of applications in protein modelling and design ${ }^{2-6}$. Here we describe our AlphaFold 3 model with a substantially updated diffusion-based architecture that is capable of predicting the joint structure of complexes including proteins, nucleic acids, small molecules, ions and modified residues. The new AlphaFold model demonstrates substantially improved accuracy over many previous specialized tools: far greater accuracy for protein-ligand interactions compared with state-of-the-art docking tools, much higher accuracy for protein-nucleic acid interactions compared with nucleic-acid-specific predictors and substantially higher antibody-antigen prediction accuracy compared with AlphaFold-Multimer v.2.3 ${ }^{7,8}$. Together, these results show that high-accuracy modelling across biomolecular space is possible within a single unified deep-learning framework.
Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41586-024-07487-w.
@article{Jumper2021, title = {Highly accurate protein structure prediction with AlphaFold}, author = {Jumper, J. et al.}, journal = {Nature}, volume = {596}, pages = {583--589}, year = {2021}, }
@article{Kreitz2023, title = {Programmable protein delivery with a bacterial contractile injection system}, author = {Kreitz, J. et al.}, journal = {Nature}, volume = {616}, pages = {357--364}, year = {2023}, }
@article{Lim2023, title = {In silico protein interaction screening uncovers DONSON's role in replication initiation}, author = {Lim, Y. et al.}, journal = {Science}, volume = {381}, pages = {eadi3448}, year = {2023}, }
@article{Mosalaganti2022, title = {AI-based structure prediction empowers integrative structural analysis of human nuclear pores}, author = {Mosalaganti, S. et al.}, journal = {Science}, volume = {376}, pages = {eabm9506}, year = {2022}, }
@misc{Anand2022, title = {Protein structure and sequence generation with equivariant denoising diffusion probabilistic models}, author = {Anand, N. and Achim, T.}, year = {2022}, eprint = {arXiv:2205.15019}, }
@article{Yang2023, title = {AlphaFold2 and its applications in the fields of biology and medicine}, author = {Yang, Z. and Zeng, X. and Zhao, Y. and Chen, R.}, journal = {Signal Transduct. Target. Ther.}, volume = {8}, pages = {115}, year = {2023}, }
@misc{Evans2022, title = {Protein complex prediction with AlphaFold-Multimer}, author = {Evans, R. et al.}, year = {2022}, eprint = {bioRxiv:2021.10.04.463034}, }
@misc{Zidek2022, title = {AlphaFold v.2.3.0 Technical Note}, author = {Židek, A.}, year = {2022}, url = {https://github.com/google-deepmind/alphafold/blob/main/docs/technicalnotev2.3.0.md}, }
@article{Isert2023, title = {Structure-based drug design with geometric deep learning}, author = {Isert, C. and Atz, K. and Schneider, G.}, journal = {Curr. Opin. Struct. Biol.}, volume = {79}, pages = {102548}, year = {2023}, }
@article{Lin2023, title = {Evolutionary-scale prediction of atomic-level protein structure with a language model}, author = {Lin, Z. et al.}, journal = {Science}, volume = {379}, pages = {1123--1130}, year = {2023}, }
@misc{Baek2021, title = {Accurate prediction of protein structures and interactions using a three-track neural network}, author = {Baek, M. et al.}, year = {2021}, eprint = {arXiv:2106.10270}, }
@misc{Wu2022, title = {High-resolution de novo structure prediction from primary sequence}, author = {Wu, R. et al.}, year = {2022}, eprint = {bioRxiv:2022.07.21.500999}, }
@article{Bryant2022, title = {Improved prediction of protein-protein interaction using AlphaFold2}, author = {Bryant, P. and Pozzati, G. and Elofsson, A.}, journal = {Nat. Commun.}, volume = {13}, pages = {1265}, year = {2022}, }
@misc{Moriwaki2021, title = {Post on X}, author = {Moriwaki, Y.}, year = {2021}, url = {https://x.com/Ag_smith/status/1417063635000598528?lang=en-GB}, }
@misc{Baek2021X, title = {Post on X}, author = {Baek, M.}, year = {2021}, url = {https://x.com/minkbaek/status/1417538291709071362?lang=en}, }
@article{Qiao2024, title = {State-specific protein-ligand complex structure prediction with a multiscale deep generative model}, author = {Qiao, Z. et al.}, journal = {Nat. Mach. Intell.}, volume = {6}, pages = {195--208}, year = {2024}, }
@article{Jumper2021, title = {Highly accurate protein structure prediction with AlphaFold}, author = {Jumper, J. et al.}, journal = {Nature}, volume = {596}, pages = {583--589}, year = {2021}, doi = {10.1038/s41586-021-03819-2}, }
@article{Kreitz2023, title = {Programmable protein delivery with a bacterial contractile injection system}, author = {Kreitz, J. et al.}, journal = {Nature}, volume = {616}, pages = {357--364}, year = {2023}, doi = {10.1038/s41586-023-05784-8}, }
@article{Lim2023, title = {In silico protein interaction screening uncovers DONSON's role in replication initiation}, author = {Lim, Y. et al.}, journal = {Science}, volume = {381}, pages = {eadi3448}, year = {2023}, doi = {10.1126/science.adi3448}, }
@article{Mosalaganti2022, title = {AI-based structure prediction empowers integrative structural analysis of human nuclear pores}, author = {Mosalaganti, S. et al.}, journal = {Science}, volume = {376}, pages = {eabm9506}, year = {2022}, doi = {10.1126/science.abm9506}, }
@misc{Anand2022, title = {Protein structure and sequence generation with equivariant denoising diffusion probabilistic models}, author = {Anand, N. and Achim, T.}, year = {2022}, eprint = {2205.15019}, archivePrefix = {arXiv}, primaryClass = {q-bio.BM}, }
@article{Yang2023, title = {AlphaFold2 and its applications in the fields of biology and medicine}, author = {Yang, Z. and Zeng, X. and Zhao, Y. and Chen, R.}, journal = {Signal Transduct. Target. Ther.}, volume = {8}, pages = {115}, year = {2023}, doi = {10.1038/s41392-023-01224-7}, }
@misc{Evans2022, title = {Protein complex prediction with AlphaFold-Multimer}, author = {Evans, R. et al.}, year = {2022}, eprint = {2021.10.04.463034}, archivePrefix = {bioRxiv}, doi = {10.1101/2021.10.04.463034}, }
@misc{Zidek2022, title = {AlphaFold v.2.3.0 Technical Note}, author = {Židek, A.}, year = {2022}, url = {https://github.com/deepmind/alphafold}, }
@article{Isert2023, title = {Structure-based drug design with geometric deep learning}, author = {Isert, C. and Atz, K. and Schneider, G.}, journal = {Curr. Opin. Struct. Biol.}, volume = {79}, pages = {102548}, year = {2023}, doi = {10.1016/j.sbi.2022.102548}, }
@article{Lin2023, title = {Evolutionary-scale prediction of atomic-level protein structure with a language model}, author = {Lin, Z. et al.}, journal = {Science}, volume = {379}, pages = {1123--1130}, year = {2023}, doi = {10.1126/science.abn2100}, }