- Course overview
- Search within this course
- An introductory guide to AlphaFold’s strengths and limitations
- Validation and impact
- Inputs and outputs
- Advanced modelling and applications of predicted protein structures
- Classifying the effects of missense variants using AlphaMissense
- Future directions and summary
- Your feedback
- Glossary of terms
- References
- Acknowledgements
All materials are free cultural works licensed under a Creative Commons Attribution 4.0 International (CC BY 4.0) license, except where further licensing details are provided.
Accessing predicted protein structures in the AlphaFold Database
Users can download many predicted protein structures from the AlphaFold Protein Structure Database (AFDB), without running the AlphaFold algorithm. Each predicted structure comes with valuable metadata and confidence metrics, supporting critical assessment. The database is freely available and can be accessed via multiple routes.
Contents of the AFDB
Google DeepMind and EMBL-EBI collaboratively developed the AlphaFold Protein Structure Database (AFDB). Their aim was to democratise predictions and make them widely available to the scientific community.
As of 2023, the AlphaFold Protein Structure Database (AFDB) hosts over 214 million entries: equivalent to almost the entirety of UniProt. The collection includes protein structures from over 1 million organisms, including model organisms and WHO pathogens of interest.
Users can search by the protein name, UniProt accession number or protein sequence.
Every predicted protein structure has a dedicated page. This contains a visualisation of the structure, plus instant access to the pLDDT scores and PAE plot. This allows fast judgement of the prediction quality for the individual structures. The structure page also contains related links and information, including a link to the protein’s UniProt page and information on the protein’s function and structural cluster.
Users can download data from AFDB. The atomic coordinates are available in PDB and mmCIF formats, while PAEs are presented in JSON format.
The data files in the archive are versioned, and previous versions are available via FTP, but the web pages will always display the latest version.
