AlphaFold Database Offers Open Access to 200 Million Predicted Protein Structures
The AlphaFold Protein Structure Database, a collaboration between Google DeepMind and EMBL-EBI, provides free access to over 200 million AI-generated protein structure predictions. This massive open-access resource aims to accelerate biological research and improve global health.
The AlphaFold Protein Structure Database provides open access to over 200 million protein structure predictions, dramatically accelerating scientific research worldwide. Developed through a partnership between Google DeepMind and EMBL's European Bioinformatics Institute (EMBL-EBI), this resource leverages advanced artificial intelligence to map the 3D structures of proteins directly from their amino acid sequences. The database achieves a level of accuracy that competes directly with traditional experimental methods.
This massive repository offers broad coverage of UniProt, the standard repository for protein sequences and annotations. Researchers easily download predicted structures for the entire human proteome, the manually curated Swiss-Prot subset, and the proteomes of 47 other organisms that play critical roles in global health and scientific research. Examples of these accessible structures include proteins linked to malaria parasite immune defense and plant disease resistance.
While the AlphaFold system still faces some limitations, its top-ranked performance in the CASP14 competition proves its immediate potential to advance biological understanding. For research needs that extend beyond the current database, scientists utilize the available open-source code to generate their own custom predictions. EMBL-EBI, Google DeepMind, NVIDIA, and Seoul National University also continue to expand the platform's capabilities, with recent updates focusing on predicting protein complexes.