AlphaFold is an artificial-intelligence system developed by Google DeepMind for predicting the three-dimensional structure of proteins from amino-acid sequences. It is regarded as one of the most important AI breakthroughs in the life sciences: AlphaFold 2 (2020) won the CASP14 competition with accuracy comparable to experimental methods, and the results were published in Nature in July 2021. For protein structure prediction, Demis Hassabis and John Jumper were awarded half of the 2024 Nobel Prize in Chemistry.
As a finished product, AlphaFold is offered in two main forms. The AlphaFold Protein Structure Database (AlphaFold DB), run jointly with EMBL-EBI and launched on July 22, 2021, contains over 200 million predicted protein structures (since 2022) and is freely available under a CC-BY 4.0 license. AlphaFold Server (2024) is a free web tool that lets researchers run AlphaFold 3 predictions for non-commercial use.
AlphaFold 3, announced on May 8, 2024 and co-developed with Isomorphic Labs, extends capabilities beyond single proteins - it predicts the structures of protein complexes with DNA, RNA, ligands and ions, with a substantial accuracy improvement for protein-molecule interactions. The system is widely used in structural biology, bioinformatics, antibody research and drug discovery.
