Google Releases 'AlphaGenome Atlas' to Predict Effects of All Single-Base Variants in the Human Genome

Google DeepMind and Google Cloud have released AlphaGenome Atlas, which predicts the effects of all possible single nucleotide variants (SNVs) in the human genome. This database utilizes the AlphaGenome AI model to pre-compute the regulatory impacts of approximately 9 billion genetic changes on molecular biology, comprising a total of 1-petabyte.

Interpretation of Non-Coding Regions and AVI Score

The protein-coding regions of the human genome account for only 2% of the total, while the remaining 98%, the non-coding regions, have been relatively poorly understood. AlphaGenome Atlas predicts how single changes in this non-coding DNA disrupt molecular processes such as protein production. In particular, it introduces the AlphaGenome Variant Impact (AVI) score, which combines predictions for coding and non-coding regions, enabling researchers to quickly prioritize the most promising research directions without manually reviewing thousands of data points.

Real-World Research Outcomes and Accessibility

This tool is achieving significant results in actual research. Laura Covill's team at the Broad Institute used the AVI score to discover a lethal variant in the DNM1 gene during research on unsolved rare diseases. Additionally, Dr. Gareth Hawkes analyzed data from over 54,000 individuals in the UK Biobank, discovering an additional 22% of non-coding genetic associations and identifying 19 genetic regions related to BMI. AlphaGenome Atlas has been opened for immediate use by clinical researchers and biologists worldwide through an intuitive interface that requires no coding skills.