The development
What happened
Google DeepMind released AlphaGenome Atlas on 8 September, making precomputed predictions for about nine billion possible single-letter DNA changes searchable. Its Variant Impact score helps rank variants for follow-up research. These are model predictions, not laboratory measurements or established clinical interpretations.
Key facts
Google DeepMind introduced AlphaGenome Atlas on 8 September 2026, precomputing AlphaGenome predictions for about nine billion possible single-letter DNA changes across the human genome.
DeepMind characterizes Atlas as a collection of model predictions rather than nine billion laboratory measurements or established clinical interpretations, totaling roughly a petabyte of data.
The release includes an AlphaGenome Variant Impact (AVI) score combining AlphaGenome and AlphaMissense predictions to rank variants, linked to feature attributions and other predictions.
DeepMind says academic researchers can use a website portal, the AlphaGenome API and an Antigravity skill, describing non-commercial access at launch and commercial access on Google Cloud as forthcoming.
DeepMind reports collaborator work in rare disease and population research, including experimental checks of selected variants. It says Atlas information is not medical advice and AlphaGenome is not approved for clinical use.
The context
Why it matters
Researchers can search and rank a genome-scale set of predictions without running each variant separately, helping them choose follow-up experiments. Selected experimental checks do not validate every prediction. Atlas information is not medical advice, and the AlphaGenome model is not validated or approved for clinical use.