GoKawiilDeepMind releases AlphaGenome 2 to predict DNA effects at single-cell level
Google DeepMind has released AlphaGenome 2, a successor to its AlphaGenome genome-prediction model, trained on more than 100 million individual cell profiles across 68 tissues. The model predicts how genetic variants affect specific cell types and states rather than cells broadly, and DeepMind says it identifies nearly twice as many expression-changing variants as its prior best model.
GoKawiil's interpretation of the reporting above, not reported fact.
By linking genetic variants to behavior in specific cell types rather than cells generally, the model could help researchers pinpoint which cells are implicated in particular diseases, according to DeepMind. The company's claim that the model generalizes to new cell types without retraining suggests it could be applied broadly across existing cell atlas data, though real-world validation in disease research will determine its practical impact.
- AlphaGenome 2 predicts genetic variant effects at the level of individual cell types and states, not just cells broadly.
- It was trained on over 100 million cell profiles spanning 68 tissues and reportedly finds nearly twice as many expression-altering variants as the prior model.
- DeepMind says the model can make predictions about new cell types without retraining, potentially extending its use across existing cell atlas datasets.
Source: theverge.com — Robert Hart, 2026-11-11
Published there as: “Google’s AlphaGenome 2 zooms in on the cellular roots of disease”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.