Skip to content
NewA calmer, faster ZeroHour
StoryModel releaseinfo

DeepMind unveils AlphaGenome 2 for cell-level DNA predictions

2 sources · 2 articlesFirst reported Updated ()
What's new

First merged summary.

Merged summary · x-ai/grok-4.7 · rewritten as coverage arrivesSummaries by Vani AI

Google DeepMind released AlphaGenome 2, which predicts how DNA variants affect cells, via a non-commercial API.

Google DeepMind unveiled AlphaGenome 2 on November 11, 2026, a conditional successor to AlphaGenome that predicts how a DNA sequence and its variants affect one tissue, cell type, or cell state. The company says it was trained on more than 100 million single-cell profiles across 68 tissues, finds nearly twice as many expression-changing variants as the prior model, and can predict unseen cell types without retraining. Outputs now include DNA methylation; inputs remain one-million-letter sequences paired with a biological condition. Non-commercial research is open through the AlphaGenome API, while commercial research is in a limited private preview. The original AlphaGenome remains for cheaper genome-wide predictions, and DeepMind says finer predictions cost more to compute.

Key facts
  • Trained on more than 100 million profiles across 68 tissues.
  • Nearly twice as many expression-changing variants as the prior model.
  • TenK10K immune-cell causal-variant recall reported at 30% versus 16%.
  • Non-commercial API access; commercial research in private preview.
  • Original AlphaGenome still handles cheaper genome-wide predictions.
  • DeepMind plans more data and possible multi-species training.
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

Timeline

dated events in the reporting · oldest first
  1. Nov 11, 2026DeepMind unveils AlphaGenome 2 and opens research API · Google DeepMind

What each outlet adds

beyond the shared facts
  • The VergeOriginal model stays; unseen cell types and multi-species plans.
  • Google DeepMindAdds methylation, TenK10K recall, and cell-encoder conditioning.

Coverage

oldest first · one row per article · UTC
  1. · 5h ago The Verge · AIFirst report

    DeepMind unveils AlphaGenome 2 for cell-level genetic variant effects

    Trained on more than 100 million cell profiles, it is available today for non-commercial research through the AlphaGenome API.

  2. · 5h ago Google DeepMind

    DeepMind announces AlphaGenome 2 for cell-type DNA predictions

    The conditional model, trained on over 100 million cell profiles, is available for non-commercial use via an API.