AlphaFold2 is assessed blind at CASP14

Assembled from 38 blocks · 1 evidence · 43 related

1. Story
2. At the fourteenth Critical Assessment of Structure Prediction, AlphaFold2 produced predictions against sequences whose experimentally determined structures had been solved but withheld, and was scored by independent assessors. The median domain GDT_TS is reported at 92.4, including on the free-modelling targets where no related structure was available to copy from — a level the assessment community described as competitive with experimental determination for most targets. The evidential design is the point: the claimant did not choose the targets, could not have trained on the answers, and did not do the scoring.
3. Knowledge
4. CASP — Critical Assessment of Structure Prediction
5. AlphaFold
6. John Jumper
7. Google DeepMind
8. AlphaFold2 is assessed blind at CASP14
9. Connections
10. AlphaFold
11. CASP — Critical Assessment of Structure Prediction
12. John Jumper
13. Google DeepMind
14. CASP — Critical Assessment of Structure Prediction
15. Google DeepMind
16. John Jumper
17. Deep learning
18. AlphaFold Protein Structure Database
19. AlphaFold2 is assessed blind at CASP14
20. AlphaFold2 is published and the database opens
21. The AlphaFold database expands to over 200 million structures
22. The Nobel Prize in Chemistry recognises protein structure prediction
23. AlphaFold
24. The difficulty of evaluating machine learning systems
25. AlphaFold2 is assessed blind at CASP14
26. AlphaFold
27. AlphaFold2 is assessed blind at CASP14
28. AlphaFold2 is published and the database opens
29. The Nobel Prize in Chemistry recognises protein structure prediction
30. AlphaFold
31. GNoME and the materials-discovery claim
32. London
33. AlphaGo defeats Lee Sedol in Seoul
34. AlphaFold2 is assessed blind at CASP14
35. GNoME is announced as 2.2 million new crystal structures
36. GenCast is published with a like-for-like comparison against an operational ensemble
37. Evidence
38. AlphaFold2 predicts protein structure from sequence at accuracy competitive with experimental determination for most targets, as assessed blind at CASP14 in 2020, where predictions are reported to have achieved a median domain GDT_TS of 92.4 including on free-modelling targets. V55 verification basis: the paper was NOT fetched — nature.com is blocked to this session — but search results carried the title, journal, volume 596 and pages 583–589 and attributed them to this paper, and carried the CASP14 accuracy characterisation. The 92.4 median GDT_TS figure is attributed in retrieved summaries to the CASP14 assessment literature rather than to this paper, and a curator should confirm which document reports it before it is quoted as the paper's own statistic.