Google DeepMind Releases AlphaGenome Atlas With 9 Billion DNA Variant Predictions

Google DeepMind released AlphaGenome Atlas, a searchable one petabyte dataset containing AI generated molecular effect predictions for more than 9 billion possible single letter DNA changes across the human genome, the Stowers Institute for Medical Research said in a press release.

The Atlas combines predictions from the AlphaGenome model, AlphaMissense, and evolutionary conservation data. It assigns each variant an AlphaGenome Variant Impact score, which ranks predicted effects across protein coding and noncoding regions, including effects on gene expression, RNA splicing, and protein function.

Researchers from Google DeepMind, the Stowers Institute for Medical Research, the Broad Institute, the University of Exeter, Memorial Sloan Kettering Cancer Center, and Stanford University contributed scientific input and tested uses for the resource. Their work included identifying a previously overlooked noncoding variant associated with an unsolved rare disease case and studying rare variants linked to protein levels in data from more than 54,000 UK Biobank participants.

AlphaGenome Atlas is available through Google DeepMind for noncommercial research. It is intended to help researchers select variants for laboratory study and has not been validated or approved for clinical use.