Research
Progress update: Our latest AlphaFold model shows significantly improved accuracy and expands coverage beyond proteins to other biological molecules, including ligands
Since its release in 2020, AlphaFold has revolutionized how proteins and their interactions are understood. Google DeepMind and Isomorphic Labs have collaborated to develop a more powerful AI model that encompasses the full range of biologically-relevant molecules.
Today, a new generation of AlphaFold has been unveiled, capable of predicting nearly all molecules in the Protein Data Bank with atomic accuracy. This breakthrough expands the model’s accuracy in key biomolecule classes, including ligands, proteins, nucleic acids, and molecules with post-translational modifications.
Above and Beyond Protein Folding
AlphaFold has progressed from single-chain protein prediction to more complex structures, with AlphaFold2.3 setting new performance standards. The model’s predictions are freely available via the AlphaFold Protein Structure Database, benefiting millions worldwide.
Accelerating Drug Discovery
The latest model surpasses AlphaFold 2.3 in protein structure prediction, particularly in drug discovery applications such as antibody binding. It outperforms docking methods and can predict ligand-protein structures without prior structural information or reference.
New Understanding of Biology
AlphaFold’s enhanced modeling capabilities offer rapid and accurate insights into complex biological systems. One case involves CasLambda, part of the CRISPR system, demonstrating the model’s potential for therapeutic applications.
Advancing Scientific Exploration
AlphaFold’s AI prowess is poised to accelerate scientific understanding across various fields. The model’s capabilities signify a substantial leap in molecular research and scientific exploration.
Our teams at Google DeepMind and Isomorphic Labs continue to push boundaries in this vital area, with ongoing progress paving the way for groundbreaking discoveries.