EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Hacker News·Sep 8·5 min read
Google DeepMind Releases AlphaGenome Atlas
TL;DR
Google DeepMind has launched AlphaGenome Atlas, a comprehensive database that predicts the effects of all 9 billion single nucleotide variants in the human genome, enabling researchers to better understand non-coding DNA regions. The Atlas features the AlphaGenome Variant Impact score, streamlining the research process by prioritizing significant genetic changes.
✦ Why It Matters
Explore the AlphaGenome Atlas today to enhance your research on genetic variants and their impacts.
Key Takeaways
- Utilize AlphaGenome Atlas for genomic variant analysis.
- Leverage the AVI score to prioritize research variants.
- Access the Atlas without coding skills.
- Explore non-coding genetic associations for complex traits.
Full Summary
AlphaGenome Atlas provides a high-resolution map of human DNA, focusing on the largely unexplored 98% of the genome that does not code for proteins. Utilizing the AlphaGenome AI model, researchers calculated the regulatory impact of 9 billion single-letter genetic changes, resulting in a 1-petabyte dataset.
The Atlas features the AlphaGenome Variant Impact (AVI) score, which simplifies the prioritization of variants for research. For instance, at the Broad Institute, the AVI score helped identify a critical variant linked to a rare disease.
Additionally, Dr. Gareth Hawkes used the Atlas to analyze data from over 54,000 UK Biobank participants, uncovering 22% more non-coding genetic associations related to body mass index (BMI).
This tool is now accessible through an easy-to-use web portal, democratizing genomic research for scientists worldwide.