Google DeepMind releases AlphaGenome 2 for research on how DNA variants affect cells
Google DeepMind's post dated 11 November 2026 announces AlphaGenome 2, a model that predicts how DNA sequences and their variations affect specific human cell types. The post says the model identifies nearly twice as many variants that change gene expression in cells, compared with its previous state-of-the-art model. It says the model was trained on more than 100 million individual cell profiles across 68 tissues. A later section says the training data include over 107 million individual cell profiles from human and mouse samples. The post does not state a parameter count or a price.
Source: Google DeepMind (opens in a new tab), 11 Nov 2026
Google DeepMind's post dated 11 November 2026 announces AlphaGenome 2, a model that predicts how DNA sequences and their variations affect specific human cell types. The post says the model identifies nearly twice as many variants that change gene expression in cells, compared with its previous state-of-the-art model. It says the model was trained on more than 100 million individual cell profiles across 68 tissues. A later section says the training data include over 107 million individual cell profiles from human and mouse samples. The post does not state a parameter count or a price.
On the TenK10K dataset, which the post describes as a single-cell map of 28 primary immune cell types, the post says AlphaGenome 2 identified nearly twice as many causal variants (30% vs. 16%) and directional gene effects (43% vs. 23%) as the original AlphaGenome fine-tuned on single-cell data and other specialised tools. Those percentages are the post's.
The post says AlphaGenome 2 is available for non-commercial research through the AlphaGenome API, and in private preview for commercial research. It says AlphaGenome 2 Skills in Google Antigravity will follow, and it does not give a date for that. It says AlphaGenome and AlphaGenome 2 are intended for research use only and have not been validated or approved for clinical or diagnostic purposes.
What it means for business: the general release named on the page is for non-commercial research. Commercial research is described as a private preview. The page does not state a token price, a parameter count, or a clinical use. This describes the post. It is not advice on using the model in a health product.