Google DeepMind announced it has used artificial intelligence to predict the biological consequences of all 9 billion possible single-letter changes to human DNA, making the resulting database freely available to academic researchers worldwide. Called AlphaGenome Atlas, this resource enables access to a comprehensive map of human genetic variation by simply opening a browser for the first time, said Pushmeet Kohli, DeepMind’s vice president of research. Each variant is linked to an average of approximately 27,000 individual predictions about its effects on gene expression and protein manufacturing, with a summary metric called AVI (AlphaGenome Variant Impact) to assess impact. Early testing shows promising results: in a patient with epileptic encephalopathy, Atlas helped reclassify a variant as likely pathogenic, and analysis of whole-genome data from over 54,000 UK Biobank participants identified 22% more associations compared to traditional methods.
Source: Read the original article
