Releases·Tool Releasefeatured
Google DeepMind unveils AlphaGenome Atlas, a predictive map of every possible DNA change in the human genome
Google DeepMind launched AlphaGenome Atlas, an AI-generated searchable database predicting the biological impact of every possible single-nucleotide change in the human genome. Precomputing these roughly 9 billion variant predictions lets researchers look up effects without running AlphaGenome per mutation, covering non-coding DNA too. DeepMind positions it as accelerating genetic research and paving the way for new disease treatments.
6103 engagement8 sourcesTue, Sep 8, 2026, 02:00 PM
AI Overview
- Google DeepMind released AlphaGenome Atlas, described as a "predictive map of every possible DNA letter change in the human genome."
- The human genome contains roughly three billion letter pairs written in four chemical letters: A, C, G, and T.
- DeepMind scientists say the tool could accelerate research and lead to new disease treatments.
- Google DeepMind precomputed predictions for roughly 9 billion single-nucleotide changes in a searchable, roughly 1-petabyte database, eliminating per-mutation model runs.
- DeepMind introduced an AlphaGenome Variant Impact (AVI) score, combining AlphaGenome, AlphaMissense and other features, that ranks mutations and reveals damage mechanisms like broken gene switches or RNA splicing.
- The Atlas targets non-coding DNA, where researchers have used the predictions to identify a likely rare-disease mutation and links between non-coding variants and BMI.
- The predictions cover molecular effects, not disease outcomes, so findings still require experimental and clinical validation.
- DeepMind announced the launch on X on 2026-09-08, describing the Atlas as mapping all 9 billion possible single-letter DNA changes.
Entities
Related
Developments·Paper
DeepMind releases AlphaGenome Atlas, a predictive map of every possible DNA letter change in the human genome
Google DeepMind launched AlphaGenome Atlas, described as a predictive map of every possible DNA letter change in the human genome. The announcement came via DeepMind and Google blogs, with a Two Minute Papers YouTube video covering it after an invite from DeepMind. For practitioners in genomics and AI for biology, it extends the AlphaGenome line toward exhaustive single-nucleotide variant prediction.
Google DeepMind launched AlphaGenome Atlas, described as a predictive map of every possible DNA letter change in the human genome. The announcement came via DeepMind and Google blogs, with a Two Minute Papers YouTube video covering it after an invite from DeepMind. For practitioners in genomics and AI for biology, it extends the AlphaGenome line toward exhaustive single-nucleotide variant prediction.
2.2k engagement1 sourceyoutubeWed, Oct 7, 2026, 01:29 PM
Social·Community
Company touts AI models aiding disease detection, from rare genetic variants to faster drug discovery
AI lab highlights its models' roles in rare genetic variant detection and drug discovery
AI lab highlights its models' roles in rare genetic variant detection and drug discovery
761 engagement1 sourcexFri, Oct 16, 2026, 10:15 AM
News·Company News
Google details AI science push: AlphaGenome Atlas, WeatherNext 3, Planetary Prediction Engine, 300-language support
Google announced that its AI technologies now support over 300 languages spoken by 7 billion people (86% of the global population), alongside new Gemini speech models, open speech datasets, on-device translation models, and science tools. It also released the AI & Economy ATLAS interactive insights and highlighted AI advances in health, weather, and crisis prediction.
Google announced that its AI technologies now support over 300 languages spoken by 7 billion people (86% of the global population), alongside new Gemini speech models, open speech datasets, on-device translation models, and science tools. It also released the AI & Economy ATLAS interactive insights and highlighted AI advances in health, weather, and crisis prediction.
0 engagement2 sourcesrssTue, Sep 15, 2026, 04:00 PM
Products·launched
AskCyph Health Explorer — DNA mutation education tool with AlphaGenome and in-browser MedGemma
An educational web tool for exploring how single-letter DNA mutations cause disease, alter 3D anatomy, inheritance odds, and treatments. It pairs DeepMind's AlphaGenome for molecular biology predictions with MedGemma 1.5 4B running locally in the browser via WebGPU as a privacy-preserving explainer. Aimed at learners wanting interactive genetics education without sending health questions to a server.
An educational web tool for exploring how single-letter DNA mutations cause disease, alter 3D anatomy, inheritance odds, and treatments. It pairs DeepMind's AlphaGenome for molecular biology predictions with MedGemma 1.5 4B running locally in the browser via WebGPU as a privacy-preserving explainer. Aimed at learners wanting interactive genetics education without sending health questions to a server.
20 engagement1 sourceredditMon, Oct 12, 2026, 09:20 PM
Social·Community
DeepMind's AlphaGenome Atlas precomputes ~9bn single-base and ~100m indel predictions, sparking debate over clinical value
A Reddit user who read the AlphaGenome Atlas preprint from DeepMind describes it as a precomputed lookup of AlphaGenome predictions over ~9 billion possible single base changes in the human genome plus ~100 million observed indels, plus a trained score (AVI) and a derived motif map. They ask genomics practitioners to explain why it matters, questioning the value of stored model predictions rather than measurements.
A Reddit user who read the AlphaGenome Atlas preprint from DeepMind describes it as a precomputed lookup of AlphaGenome predictions over ~9 billion possible single base changes in the human genome plus ~100 million observed indels, plus a trained score (AVI) and a derived motif map. They ask genomics practitioners to explain why it matters, questioning the value of stored model predictions rather than measurements.
23 engagement1 sourceredditThu, Sep 10, 2026, 12:47 AM