AlphaMissense classifies 89% of all 71 million possible human missense mutations
Result confirmed
The takeaway
AlphaMissense (Science, Sept 2023) scored all ~71 million possible single amino-acid substitutions in 19,233 human proteins and classified 89%: 57% likely benign and 32% likely pathogenic.
Status
- Claim
Result confirmed
- Our reporting
- High confidence
- Verification
- Peer-reviewed
- Importance
- 3 of 5
- Last verified
- 29 September 2026
Your AI and this story
- GPT-6 AstraIn its training data
- Claude Opus 5.5In its training data
- Gemini 4 HydrogenIn its training data
- Grok 4.7In its training data
It happened before the cutoffs of all four assistants, so it can be in their training data.
Key facts
- 71M variants scored; 89% classified (57% likely benign, 32% likely pathogenic)
- Human experts had confidently classified only ~0.1% of missense variants
- Predictions released freely; model weights restricted
What happened
Built on AlphaFold-style protein modelling, AlphaMissense predicted which mutations likely disrupt protein function.
Why it matters
It gives clinicians a first-pass interpretation for millions of variants they otherwise could not assess.
Sources
2 sources from 2 sites. Numbers match the chips in the text.
2 sources: 2 primary
Primary
- Accurate proteome-wide missense variant effect prediction with AlphaMissense (Science)science.org, paper
- DeepMind: A catalogue of genetic mutations to help pinpoint the cause of diseasesdeepmind.google, official
Changes
- Filed