Moderate evidence
More
AlphaFold's protein database grew from 300,000 to over 214 million AI predictions
Study title: AlphaFold Protein Structure Database in 2024: providing structure coverage for over 214 million protein sequences
- Evidence type
- Official statistics
- Sample
- n = 214,000,000
- Sources
- 1 sourcePeer-reviewed
- Published
- 1 min read
- Truvace record
- TRV-2026-1205cite or check it
Study title: AlphaFold Protein Structure Database in 2024: providing structure coverage for over 214 million protein sequences
By October 2023, the authors reported that the AlphaFold Protein Structure Database had grown from 300,000 structures in 2021 to over 214 million predicted structures, enabled by AlphaFold2. The paper outlines archiving enhancements, coverage of model organisms and other curated sets, and access routes including FTP and Google Cloud Public Datasets.
Read the study abstract
The AlphaFold Database Protein Structure Database (AlphaFold DB, https://alphafold.ebi.ac.uk) has significantly impacted structural biology by amassing over 214 million predicted protein structures, expanding from the initial 300k structures released in 2021. Enabled by the groundbreaking AlphaFold2 artificial intelligence (AI) system, the predictions archived in AlphaFold DB have been integrated into primary data resources such as PDB, UniProt, Ensembl, InterPro and MobiDB. Our manuscript details subsequent enh…

The gain
The AlphaFold database's authors report it grew from 300,000 predicted protein structures in 2021 to over 214 million, with predictions now integrated into PDB, UniProt, Ensembl, InterPro and MobiDB.
In brief
The expansion matters because the predictions have been integrated into primary resources such as PDB, UniProt, Ensembl, InterPro and MobiDB, extending structural information available for biological research workflows. The source does not quantify downstream discovery outcomes or validation rates, leaving uncertainty about coverage quality across all 214 million entries.
Main points
- AlphaFold DB grew from initial 300k structures released in 2021 to over 214 million predicted structures by 2024 coverage described in 2023 manuscript.
- Predictions are enabled by the AlphaFold2 artificial intelligence system.
- Structures have been integrated into PDB, UniProt, Ensembl, InterPro and MobiDB.
- Paper details data access via FTP, Google Cloud Public Datasets, and programmatic endpoints, plus improvements to Predicted Aligned Error viewer, 3D viewer customization, and search engine.
The full story
The manuscript describes successive releases covering model organisms, global health proteomes, Swiss-Prot integration and curated protein datasets, expanding from the 2021 initial release.
Access mechanisms include direct file access via FTP, advanced queries using Google Cloud Public Datasets, and programmatic access endpoints, with added services for viewing Predicted Aligned Error and customizing the 3D viewer.
Sources
1 on file- Nucleic Acids Researchdoi:10.1093/nar/gkad1011
Talk it over
The authors report the database grew from 300,000 predicted structures in 2021 to over 214 million. A count of predictions is not a count of discoveries, so what would you look for to tell whether biologists benefit?
AlphaFold's protein database grew from 300,000 to over 214 million AI predictionshttps://truvace.com/item/alphafold-protein-structure-database-in-2024-providing-structure-coverage-for-ov
The authors report the database grew from 300,000 predicted structures in 2021 to over 214 million. A count of predictions is not a count of discoveries, so what would you look for to tell whether biologists benefit?
Answer in the threadThe predictions are now integrated into PDB, UniProt, Ensembl, InterPro and MobiDB. Who gains most from that, and who might never see the benefit?
Answer in the threadThe source does not quantify validation rates for the 214 million entries. What finding about their quality would change how much weight you give this growth?
Answer in the thread
If you have more time
Look up how the database's releases for model organisms and global health proteomes were used. What kind of evidence would show that these predictions led to discoveries?
Answer in the threadThe paper describes access through FTP, Google Cloud Public Datasets and programmatic endpoints. Which of these routes would a small lab or a school find easiest, and what would you ask to find out?
Answer in the thread
ace
What readers say