Google DeepMind has released the AlphaGenome Atlas, a free public database that predicts the molecular effects of all 9 billion possible single-letter changes to human DNA. The dataset spans more than 1 petabyte, over 30 times larger than DeepMind's AlphaFold protein database.
The Atlas predicts how each mutation affects gene regulation across hundreds of human and mouse cell types and tissues, and introduces a new AlphaGenome Variant Impact (AVI) score that ranks how significant a variant is likely to be, covering both protein-coding and non-coding regions of the genome. Testing all 9 billion variants in a lab would be practically impossible, so the computational predictions let researchers narrow down candidates fast.
Early collaborators used the tool to identify mutations linked to epileptic encephalopathy and found 22% more genetic associations in protein studies than previous methods. The Atlas is available for free via a web portal and API, with additional access through Google Cloud.