Stowers Institute Partners with Google DeepMind to Launch AlphaGenome Atlas

Researchers at the Stowers Institute for Medical Research have partnered with Google DeepMind and other leading institutions to develop a groundbreaking new resource called AlphaGenome Atlas. This one-petabyte dataset contains AI-generated predictions for more than 9 billion possible single-letter DNA changes across the human genome, providing scientists with a comprehensive, searchable map of potential genetic variations.

The project, which has taken several years to develop, aims to accelerate the understanding of the human genome. Until now, scientists have lacked a single resource that could both rank variants across the genome and reveal the biological processes they are predicted to disrupt. AlphaGenome Atlas fills this gap, offering a powerful tool for researchers to quickly prioritize and interpret genetic variants that may influence biology and disease.

Julia Zeitlinger, a Stowers Institute Investigator, played a crucial role in the project. She, along with researchers from Google DeepMind, the Broad Institute, Memorial Sloan Kettering Cancer Center, the University of Exeter, and Stanford University, worked together to map and interpret the patterns in DNA that regulate biological processes inside cells. Zeitlinger, who also leads the Stowers Institute's AI Initiative, brought her expertise in gene regulation and computational biology to the project.

"We can quickly query many cell types and look for general patterns by which genes are activated and repressed," Zeitlinger said. "This is a very difficult problem because every cell type speaks a slightly different language, making it hard to know which rules are general."

AlphaGenome Atlas contains thousands of molecular-effect predictions for each variant across hundreds of human cell types and tissues. These predictions contribute to the new AlphaGenome Variant Impact (AVI) score, which gives researchers a single measure for ranking variants by their potential impact across protein-coding and non-coding regions of the genome.

Scientists can use the AVI score to identify and map recurring DNA motifs, short sequences where transcription factors bind to help control gene activity. The resource has already been applied to genomic data from the UK Biobank, uncovering additional associations between rare noncoding variants and protein levels.

"At the Stowers Institute, we are committed to pairing deep biological knowledge with the capabilities of AI to ask questions at a scale that was not previously possible," said Alejandro Sánchez Alvarado, Stowers Institute President and Chief Scientific Officer.

The Atlas does not replace laboratory research but can help scientists determine which variants and biological mechanisms should be investigated first, focusing experimental time and resources on the most promising questions. The resource is available for non-commercial use through Google DeepMind's website, making it accessible to a broader community of researchers.

"AlphaGenome Atlas is a powerful example of how AI can expand human knowledge and advance scientific discovery," said Pushmeet Kohli, Vice President of Science and Chief Scientist at Google DeepMind. "By making this resource widely available, we hope scientists around the world can use it to better understand the language of life and what happens when individual letters in the human genome change."

The collaboration between the Stowers Institute and Google DeepMind highlights the potential of AI in transforming the field of genetics and genomics. By providing researchers with a comprehensive, searchable resource, the AlphaGenome Atlas is expected to accelerate foundational discoveries, disease research, and the search for therapeutic targets.