DeepMind mapped the impact of 9 billion DNA mutations

Google DeepMind just mapped the biological effect of every single-letter mutation in human DNA.

DeepMind announced AlphaGenome Atlas, a 1-petabyte database predicting how genetic changes alter molecular processes. Scientists understand the 2% of human DNA that codes for proteins, but the remaining 98% is mostly obscure. DeepMind used its AI model to pre-calculate all 9 billion single-letter genetic variants, boiling the results down into a single metric called the AVI score.

Why it matters: It turns months of lab filtering into a quick database search. Researchers at the Broad Institute already used the score to solve an open rare disease case by pinpointing a mutation in the DNM1 gene. In another test with 54,000 UK Biobank participants, the tool uncovered 22% more non-coding genetic associations and flagged 19 regions tied to body mass index.

You can search the database today through a web portal with zero coding required.

Finding needle-in-a-haystack mutations just got as fast as a web search.