Google Deepmind has released AlphaGenome 2, a new AI model for genomics research. It predicts the effects of genetic variants on specific human cell types and finds nearly twice as many relevant gene changes as its predecessor.
Every cell in the human body carries the same DNA blueprint, but different cell types switch different genes on or off. A heart muscle cell reads DNA differently than an immune cell. Predicting those differences is one of the biggest unsolved problems in genomics. Google Deepmind aims to take a major step in that direction with AlphaGenome 2.
The new model is designed to predict how DNA sequences and their variations affect specific human cell types. It was trained on more than 107 million individual cell profiles from 68 tissues, according to Google Deepmind, and is meant to help researchers uncover the cellular drivers of disease.
AlphaGenome 2 shifts from tissue-level overviews to individual cells
Its predecessor, AlphaGenome, took in a DNA sequence a million letters long and predicted its properties across thousands of experiments spanning many tissues at once. According to Google Deepmind, that parallel approach spread the model’s capacity too thin and made it hard to capture subtle differences between closely related cell types.
AlphaGenome 2 instead works as a conditional model. Along with the DNA sequence, it takes a specific biological state as an extra input and calculates effects for that one cell type. The system was trained on more than 107 million individual cell profiles from 68 tissues in humans and mice.
Beyond predicting how strongly a gene is expressed, the model also calculates how DNA is physically packaged in the nucleus, showing which stretches of the genome are accessible. It can now also predict DNA methylation, chemical tags that act as molecular switches to silence or activate genes.
AlphaGenome 2 can also infer missing data from a single measurement. If researchers measure only a cell’s gene activity, for instance, the model provides predictions for additional properties, such as how tightly or loosely certain stretches of DNA are packaged. That helps researchers decide which lab experiments are worth running next.
AlphaGenome 2 pinpoints disease genes and vaccine responses more accurately
On the TenK10K benchmark dataset with 28 primary immune cell types, AlphaGenome 2 identified about 30 percent of causal variants. In 43 percent of cases, it also correctly predicted whether a variant upregulates or downregulates a gene. A fine-tuned version of AlphaGenome 1 managed just 16 and 23 percent. The system also models continuous changes, like the maturation of blood stem cells into red blood cells, with considerably greater accuracy.
Working purely computationally, the model reproduced an inflammatory bowel disease mechanism previously identified in a study. It isolated the exact DNA variant in a regulatory switch far from the gene that selectively overactivates ETS2 in gut macrophages, while other cell types and genes remained largely unaffected.
A key advantage of the new architecture is its ability to adapt to entirely new experiments without retraining. Researchers can feed in gene activity data from just a few dozen cells, and the model calculates effects for that previously unseen cell state. According to Deepmind, it often beats specialized models trained directly on that data.
Applied to data from a COVID-19 mRNA vaccine study, the model reliably distinguished vaccine-activated T cells from resting ones and showed that a gene variant linked to severe COVID-19 cases disrupts cell activity mainly in the activated state. In resting cells, that effect would have remained invisible.
Deepmind’s precomputed predictions are meant to link human genetics and cell biology
Deepmind has already precomputed predictions for hundreds of cell and tissue conditions, partly to support AI-driven drug development. The precomputed predictions also help researchers analyze genome-wide association studies (GWAS). These large-scale studies link regions of DNA to traits like height or type 2 diabetes but often only roughly locate signals within broad stretches of the genome. AlphaGenome 2 narrows down the likely causal variants and identifies the affected cell types, such as astrocytes in the brain or specific immune cells. When the model is confident enough, it also predicts whether a mutation upregulates or downregulates a gene, an important capability for drug development.
AlphaGenome 2 is available now for noncommercial research through an API, while commercial users get access through a closed preview. The system is not approved for clinical diagnostics. Deepmind also plans to soon roll out dedicated AlphaGenome 2 features on its agentic development platform Google Antigravity, where the model is set to be combined with tools like Gemini 4 for automated research workflows.
AI News Without the Hype – Curated by Humans
Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive “AI Radar” frontier report six times a year, full archive access, and access to our comment section.
