Accurate structure prediction of biomolecular interactions with AlphaFold 3
By Josh Abramson, Jonas Adler, Jack Dunger, Richard Evans, Tim Green, Alexander Pritzel, Olaf Ronneberger, Lindsay Willmore, Andrew J Ballard, Joshua Bambrick and 38 more. Nature, 2024
12,322 citations recorded by the source corpus.
The introduction of AlphaFold 21 has spurred a revolution in modelling the structure of proteins and their interactions, enabling a huge range of applications in protein modelling and design2, 3, 4, 5–6. 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.2.37,8. Together, these results show that high-accuracy modelling across biomolecular space is possible within a single unified deep-learning framework. AlphaFold 3 has a substantially updated architecture that is capable of predicting the joint structure of complexes including proteins, nucleic acids, small molecules, ions and modified residues with greatly improved accuracy over many previous specialized tools.
Subfields: Molecular Biology, Computational Biology & Bioinformatics, AI for Physical Sciences
View Paper: https://doi.org/10.1038/s41586-024-07487-w
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Molecular Biology
- Siqi Sun
- Sheng Wang
- Jeremy Wohlwend
- Tommi S. Jaakkola
- Sebastian Nowozin
- Peng Xiong
- Pushmeet Kohli
- Tao Shen
In Computational Biology & Bioinformatics
- Siqi Sun
- Sheng Wang
- Peng Xiong
- Pushmeet Kohli
- Tao Shen
- Paris Perdikaris
- Tommi S. Jaakkola
- Jeremy Wohlwend
In AI for Physical Sciences
- Paris Perdikaris
- Tommi S. Jaakkola
- Victor Garcia Satorras
- G. Em Karniadakis
- Maizar Raissi
- Liu Yang
- Jian-Xun Wang
- Dongkun Zhang
Citations in the corpus
- Accurate RNA 3D structure prediction using a language model-based deep learning approach (2022, positive): “21) and RoseTTAFoldNA 20 can predict RNA complexes, their accuracy is still limited and they perform less well than RhoFold+ on single-strand RNAs.”
- Boltz-1 Democratizing Biomolecular Interaction Modeling (2024, positive): “Baselines We evaluate our performance against AlphaFold3 [Abramson et al., 2024] and Chai-1 [Chai et al., 2024], current state-of-the-art biomolecular structure prediction models that were released under an exclusive commercial license and do not have training code and pipelines available.”
- A foundation model for the Earth system (2025, positive): “Taking inspiration from recent successes of foundation models in other fields 4,5 , we first pretrain Aurora on more than one million hours of diverse Earth system data.”
- AlphaEvolve: A coding agent for scientific and algorithmic discovery (2025, positive): “The transformer architecture [100] is used in the majority of modern neural networks, ranging from LLMs to AlphaFold [1].”
- One Step Diffusion via Shortcut Models (2024, neutral): “…2020; Song et al., 2020) and flow-matching (Lipman et al., 2022; Liu et al., 2022) have seen remarkable success in modelling diverse images (Rombach et al., 2022; Esser et al., 2024), video (Ho et al., 2022; Bar-Tal et al., 2024), audio (Kong et al., 2020), and proteins (Abramson et al., 2024).”
- Scalable emulation of protein equilibrium ensembles with generative deep learning (2025, positive): “…sequencing has made it possible to acquire the protein sequences of entire genomes at low cost, while AlphaFold [1] and similar models [2–4] have built upon the decades of data accumulated in the Protein Data Bank (PDB) [5] to predict 3D protein structures that in many cases match…”
- Gaia: An AI-enabled genomic context–aware platform for protein sequence annotation (2025, unknown): “2 were generated using Alpha-Fold3 ( 56 ).”