Google DeepMind Unveils AlphaGenome Atlas to Map Every Possible Single-Letter DNA Mutation in the Human Genome
Google DeepMind has released the AlphaGenome Atlas, an AI-generated database that predicts the biological consequences of all 9 billion possible single-letter changes to human DNA. The atlas provides a comprehensive map of how each mutation affects gene regulation, including gene expression, splicing, and protein changes. To create the atlas, the administration announced that DeepMind used its AlphaGenome AI model to compute predictions for every possible nucleotide change across the human genome. The database also includes over 100 million insertions and deletions, along with a catalogue of more than 2,500 recurring DNA motifs. To help researchers prioritize variants, the atlas includes the AlphaGenome Variant Impact (AVI) score, a single metric that combines predictions from the AlphaGenome and AlphaMissense models. Researchers noted that the atlas serves as a searchable dictionary for non-coding DNA, which accounts for 98% of the genome. While the atlas is not a substitute for experimental evidence, it provides a tool to scale up access to genetic models and help diagnose rare diseases.
Sources
-
DeepMind’s new genome ‘atlas’ charts effects of all nine billion human gene mutations
Nature
-
Predicting genome-wide functional constraints with GPN-Star
Nature
-
Google DeepMind publishes AI-powered predictions for effect of all 9B mutations to human DNA
Fortune