AlphaFold Protein Structure Database

node · earth

The public database, built and hosted by EMBL's European Bioinformatics Institute at the Wellcome Genome Campus in Hinxton, that turned a method into infrastructure. It opened in July 2021 with predicted structures for the human proteome and a set of model organisms, and in late July 2022 was expanded roughly two hundredfold to cover more than 200 million sequences — effectively the whole of UniProt. A 2024 database paper reports coverage of over 214 million protein sequences. Adoption is the part that matters for this Journey: DeepMind and EMBL-EBI report more than three million researchers in over 190 countries using it by late 2025, over a million of them in low- and middle-income countries, and an independent analysis attributes to its users a substantial rise in submissions of novel experimental structures. That last finding is the useful one, because it distinguishes a tool being downloaded from a tool changing what work gets done.