Google unveils AlphaGenome Atlas mapping DNA mutation effects

Google has officially launched the AlphaGenome Atlas, an advanced database powered by artificial intelligence designed to predict the regulatory and biological effects of single nucleotide variants across the human genome. Leveraging its AlphaGenome AI model, Google pre-calculated the impacts of all 9 billion single-letter genetic variations, resulting in a massive 1-petabyte dataset. While science has developed a solid understanding of the 2% of DNA that codes for proteins, the remaining 98% non-coding genome has remained largely unmapped. AlphaGenome Atlas bridges this gap by introducing the AlphaGenome Variant Impact score, enabling researchers to efficiently evaluate both coding and non-coding regions without manually analyzing thousands of data points. Early research applications demonstrate substantial value: researchers at the Broad Institute used the score to prioritize rare disease variants, while analysis of over 54,000 UK Biobank participants revealed 22% more non-coding associations and identified 19 genetic regions tied to body mass index. Google made the portal accessible to the broader scientific community without requiring coding skills, reinforcing its AI leadership in life sciences.