AlphaGenome Atlas: a high-resolution map of human DNA

Google DeepMind has introduced AlphaGenome Atlas, a database that predicts the effects of every possible single nucleotide variant in the human genome.

What the Atlas contains

The human genome is made of about 3 billion base pairs of DNA. Scientists understand the roughly 2% that codes for proteins relatively well, but their knowledge of the remaining 98% is limited. The AlphaGenome model had already shown how single changes in those non-coding regions can disrupt molecular processes such as protein production, but the wider picture remained unclear.

Using the AlphaGenome AI model, the team pre-calculated the regulatory impact of all 9 billion single-letter genetic changes. The result is a one-petabyte dataset that researchers can query through the new Atlas.

A single score for prioritising variants

To make that volume of data navigable, the Atlas introduces the AlphaGenome Variant Impact (AVI) score. It combines predictions for coding and non-coding regions in one figure, letting researchers prioritise the most promising avenues without sifting through thousands of data points.

Early research uses

At the Broad Institute, Laura Covill and her team used the AVI score to prioritise variants for unsolved rare disease research. The tool highlighted a critical variant in the DNM1 gene and predicted that it created an incorrect splice site; that provided crucial supporting evidence for solving the case.

Dr. Gareth Hawkes applied the Atlas to data from more than 54,000 UK Biobank participants. Grouping variants by predicted molecular effect uncovered 22% more non-coding genetic associations, while focusing on the top 1% of impactful variants identified 19 genetic regions linked to body mass index.

Access

AlphaGenome Atlas is available today through a web portal that requires no coding skills, opening it to clinical researchers and biologists worldwide.