Google DeepMind’s AlphaGenome 2 Predicts Cell-State Variant Effects

Google DeepMind announced AlphaGenome 2 on November 11, 2026, a deep-learning model that predicts how DNA variants affect gene regulation in individual cell types and states. Trained on more than 100 million individual cell profiles, it is available for non-commercial research and in private preview for commercial research.

The release addresses a blind spot in tissue-wide genomic models: a variant can matter only in a specialized cell or in a cell activated by disease, leaving its signal diluted in a tissue average. AlphaGenome 2 instead conditions each prediction on a selected tissue, cell type or cell state, including conditions generated from a researcher’s own single-cell experiment.

The earlier AlphaGenome 1 predicted molecular activity across broad tissues, while AlphaGenome Atlas offered predictions for all 9 billion possible single-letter changes in the human genome. AlphaGenome 2 is a separate model release, not an extension of that precomputed atlas.

A Model Conditioned on Cell Identity

AlphaGenome 1 generated fixed outputs for bulk tissues and cell lines simultaneously from a one-million-letter DNA sequence. AlphaGenome 2 evaluates one condition at a time and supplies that condition throughout the network, allowing it to represent how the same DNA segment is read differently across cell types. Its readouts include gene expression, RNA splicing and chromatin accessibility, with CpG DNA methylation added to the original model’s capabilities.

The technical paper describes training on human and mouse bulk and single-cell data. Researchers can supply single-cell RNA-seq data, a snapshot of gene activity, from a new experiment, and a cell encoder turns it into a condition without further fine-tuning. That allows predictions for datasets and cell states the model has not seen before.

The same design can fill in missing experimental readouts. Given gene-expression data alone, AlphaGenome 2 can predict unmeasured properties such as chromatin accessibility—whether stretches of DNA are open to regulation—helping researchers prioritize subsequent experiments. The trade-off is computational: every condition requires a separate prediction, making the model more intensive to run than fixed-track multitask models.

Variant Benchmarks and Disease Examples

DeepMind’s paper reports that on TenK10K, a single-cell immune dataset spanning 28 cell types, AlphaGenome 2 recovered 30% of putatively causal expression-linked variants at 90% precision, compared with 16% for AlphaGenome 1 fine-tuned on single-cell data. It recovered 43% of variant effect directions at 90% accuracy, versus 23%. On genome regions excluded from training, the model also explained a median 34% of within-tissue, cell-type-specific gene-expression variance, against 23% for the same fine-tuned predecessor. These evaluations measure regulatory differences between related cells, the problem the conditional architecture is designed to address.

DeepMind also reports two research examples of cell-state-specific analysis. In an inflammatory bowel disease analysis, AlphaGenome 2 flagged rs2836882 from a seven-variant credible set, linked it to ETS2 approximately 290 kilobases away and localized the predicted effect to gut monocytes and macrophages, reproducing a mechanism previously established experimentally. Using single-cell data from a COVID-19 mRNA vaccine study that was new to the model, it distinguished vaccine-activated T cells from resting T cells and predicted a stronger effect for a variant associated with severe COVID-19 in the activated state. Together, the examples illustrate how a variant’s predicted effect can depend on both cell type and current state.

Precomputed Resources and API Access

DeepMind precomputed two resources for researchers who do not want to run large-scale comparisons themselves. One maps enhancer–gene links across 640 primary single-cell reference conditions and 39,323 protein-coding and long non-coding RNA genes; enhancers are regulatory DNA segments that can act far from their target genes. Another scores more than 6.8 million unique common and trait-associated single-letter variants across 510 human conditions to nominate possible causal variants, target genes, effect directions and relevant cell states.

At launch, AlphaGenome 2 and expression-conditioning Colab tutorials were available through the AlphaGenome API for non-commercial research, with commercial research in private preview. The official documentation says non-commercial API use is free and query rates vary with demand. DeepMind also plans AlphaGenome 2 Skills for Google Antigravity, its agentic development platform. Genomics PLC CEO Peter Donnelly said he looked forward to incorporating AlphaGenome 2 into the company’s Mystra platform and described sequence-prediction models as tools for moving from association signals to information drug-development programs need.

The paper cautions that even at high precision, AlphaGenome 2 recovers only a minority of the expression-linked and disease-associated variants that statistical analyses nominate as likely causal. It can underestimate effects outside accessible chromatin, at very distant regulatory elements or through mechanisms acting after transcription. Transient, stimulated and disease states remain underrepresented in training and the precomputed variant panel, and fine-grained neuronal subpopulations remain difficult to distinguish.

Those condition representations come from observational, steady-state profiles and therefore capture correlations rather than causal relationships. It remains uncertain whether the model can predict responses to gene knockouts, overexpression or dynamic cell-state transitions. DeepMind restricts both AlphaGenome versions to research use; neither has been validated or approved for clinical or diagnostic purposes.

The central consequence is a more targeted route from a disease association to a candidate variant, target gene and cell state. AlphaGenome 2 can narrow that search across hundreds of conditions, but its incomplete variant recovery and correlational view of cell states mean the resulting hypotheses still depend on experimental follow-up.