DeepMind releases AlphaGenome 2: it predicts variant effects per cell type and finds nearly twice as many causal variants as its predecessor (30% vs 16%)
A DNA sequence model that moves from tissue averages to individual cell types and states. What changed in the architecture, how much better it is, and how to access it.
// Key points
- It is now a conditional model: a million-letter DNA sequence plus one target condition (tissue, cell type or state), picked from a reference catalog or generated on the fly from a scientist's own single-cell experiment, without retraining.
- It is trained on bulk and single-cell data from human and mouse, including over 107 million individual cell profiles across 68 tissues, and adds DNA methylation as a readout.
- On TenK10K, a single-cell map of 28 immune cell types, it identifies 30% of causal variants versus 16%, and 43% of directional gene effects versus 23%.
- It is available today for non-commercial research via the AlphaGenome API, and in private preview for commercial research.
Builder's takeThis is far from my products; I include it for the design idea: make the target a conditional input instead of producing a fixed output per case. Products with many tenants or roles, such as AI Interview giving feedback per job, can try one model plus a condition before maintaining a prompt per role.
// Background · from #Google
Full timeline →- Nov 12 Google launches the Gemini Decisions API: Gemini Decision 1 answers up to 1,000 closed-set questions in one pass at $0.10 per million input tokens, with no output billed
- Nov 11 Gemini 4 Carbon opens to all developers at Argon's introductory price of $2 input and $10 output per million tokens
- Nov 11 DeepMind introduces AlphaProof Code: agents generate over six million lines of Lean to formally verify Signal's protocol implementation in four weeks