AlphaFold 1
Era 4 · Deep learning and foundation models (2012–2021) › AI for science
- Lineage
- Acting in the world
- Kind
- Application
Who
DeepMind
What happened
DeepMind's AlphaFold takes first place overall at CASP13, the biennial protein structure prediction competition, with the best prediction for 25 of the 43 hardest targets.
Why it mattered
Deep learning's first standout result in protein structure prediction: a network estimates the distances between amino acids and uses them to guide the folding.
The full story
A protein is a chain of amino acids that folds up on itself, and its three-dimensional shape determines what it does in the cell. Reading the sequence is cheap, but determining the shape with experimental methods such as crystallography can take months or years. Every two years the CASP competition tests the programs that try to predict it, using newly solved structures that no one knows in advance.
At CASP13, in 2018, DeepMind entered with AlphaFold. The program used a deep neural network to estimate a probability distribution over the distances between every pair of amino acids, instead of the simple contact map other methods relied on. Those distances were turned into a potential that, combined with a physical potential, guided the search for the structure.
In the hardest category, proteins with no known similar structures, it gave the best prediction for 25 of the 43 targets, with a median GDT score of 58.9, against 52.5 and 52.4 for the next two teams. It took first place overall.
The paper appeared in Nature in January 2020, with nineteen authors led by Andrew Senior. By then the team was already working on a completely redesigned system.
Leads to
AlphaFold 2 (November 2020)
Inherited, with no proven direct link
The same DeepMind team; the AlphaFold 2 paper describes it as a completely different model from the CASP13 system
Threads that pass through here
- From protein to Nobel (step 2 of 4)
Sources
Original source
- Improved protein structure prediction using potentials from deep learning. Nature, 577(7792), pp. 706-710 (opens in a new tab). Andrew W. Senior, Richard Evans, John Jumper et al., 2020
Further reading
How to cite
“AlphaFold 1.” The History of AI, revised November 13, 2026. https://thehistoryofai.com/en/milestone/alphafold-1
This entry was last reviewed on November 13, 2026.