Google DeepMind launches AlphaGenome 2 for cell-type-specific disease prediction
·2026.11.18 07:00
Key point
The new model addresses the limitation of previous tissue-level averages by identifying cellular drivers of disease, building on the AlphaGenome Atlas dataset of 9 billion single-letter changes.
Details
- Google DeepMind released AlphaGenome 2, a model designed to predict how genetic variants disrupt individual cell types rather than relying on tissue-level averages.
- The tool builds on AlphaGenome, which processes 1 million base pair sequences to make thousands of predictions, and AlphaGenome Atlas, a dataset covering 9 billion possible single-letter DNA changes.
- Researchers use the AlphaGenome Variant Impact (AVI) score to prioritize variants; for example, a University of Exeter fellow used it to narrow 1 billion variants to 19 groups linked to body mass index.
- The long-term goal is to create a fully functioning digital simulation of a human cell to model disease mechanisms and test candidate therapies.
- AlphaGenome 2 is intended to help researchers pinpoint cellular drivers of disease, while current models still require interpretation alongside other information sources due to varying confidence levels in predictions.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.