# Google DeepMind launches AlphaGenome Atlas - Dealroom

*Источник: Dealroom*
*Дата: 2026-09-08*
*Язык: en*

**Кратко:** On 8 September 2026, Google DeepMind introduced AlphaGenome Atlas, a free-to-use resource for academic research that contains predictions for the molecular effects of more than 9 billion possible single-nucleotide variants — every possible single-letter change in the human genome. The atlas precomputes AlphaGenome predictions at genome scale and adds an AlphaGenome Variant Impact (AVI) score, which combines signals from AlphaGenome and AlphaMissense to help researchers rank variants and investigate their likely biological effects.

What's the deal? On 8 September 2026, Google DeepMind introduced AlphaGenome Atlas, a free-to-use resource for academic research that contains predictions for the molecular effects of more than 9 billion possible single-nucleotide variants — every possible single-letter change in the human genome. The atlas precomputes AlphaGenome predictions at genome scale and adds an AlphaGenome Variant Impact (AVI) score, which combines signals from AlphaGenome and AlphaMissense to help researchers rank variants and investigate their likely biological effects.
Why it matters: AlphaGenome Atlas is designed to make large-scale genomic interpretation more accessible. It includes thousands of molecular-effect predictions for each variant across hundreds of human and mouse cell types and tissues, alongside feature attributions covering processes such as chromatin accessibility, RNA splicing and gene expression. The resource also includes more than 2,500 recurring DNA sequence motifs and links predicted effects to the regulatory sequences they may disrupt. Its coverage spans both coding regions, which represent about 2% of the genome, and non-coding regions, which contain most trait-associated variants.
The dataset is approximately 1 petabyte, making it more than 30 times larger than the AlphaFold Database. In early applications, collaborators used the AVI score to prioritise a variant affecting DNM1 in unsolved rare-disease research; experimental screens validated the predicted splice-site impact. Separately, analysis of whole-genome data from more than 54,000 UK Biobank participants identified 22% more non-coding genetic associations than an approach without the atlas’ predicted molecular effects. AlphaGenome Atlas is available through a web portal and API for non-commercial use; the underlying AlphaGenome model is not validated or approved for clinical use.
Read more: Google DeepMind · AlphaGenome Atlas

[Оригинал](https://app.dealroom.co/news/note/google-deepmind-launches-alphagenome-atlas)