Google Deepmind’s Alphagenome 2 Maps How Mutations Affect Cells
WHO:
The Verge - Google DeepMind researchers developed AlphaGenome 2, a successor to the AlphaGenome DNA-prediction model.
The Verge - Adam Gayoso, a Google DeepMind research scientist involved in the project, said the model’s breakthrough was “tying human genetics to cell biology.”
The Verge - Žiga Avsec, Google DeepMind’s genomics lead, said cell-specific detail matters when developing drugs that target “specific proteins and specific cell types.”
The Verge - Pushmeet Kohli heads DeepMind’s AI for science research.
WHAT:
The Verge - AlphaGenome 2 predicts regulatory activity and the effects of genetic variants in individual cell types and biological states.
The Verge - The model was trained on more than 100 million individual cell profiles across 68 tissues.
The Verge - Google said AlphaGenome 2 identifies nearly twice as many variants that can change gene expression in cells as its previous state-of-the-art model.
The Verge - The model can make predictions for new cell types and biological states without retraining, and can identify gaps in experimental data.
The Verge - AlphaGenome 2 is available through the AlphaGenome API for non-commercial research; commercial research has a limited “private preview.”
The Verge - AlphaGenome 2 does not replace AlphaGenome; its more detailed predictions are more computationally expensive.
The Verge - DeepMind’s September genome-wide Atlas covered “every possible DNA letter change in the human genome.”
MarkTechPost - AlphaGenome 2 is a conditional DNA sequence model that predicts regulatory activity and variant effects per cell type.
WHERE:
The Verge - The model’s training data covered 68 tissues; the article does not specify their locations.
The Verge - DeepMind was reported to be expected to open an automated materials science lab in the UK.
WHY:
The Verge - The model is intended to help researchers examine how genetic variants affect specific cells and connect those effects to disease, potentially informing new treatments.
The Verge - Different cell types switch different genes on and off, so the same genetic variant can have different effects depending on a cell’s type and state.
HOW:
The Verge - AlphaGenome 2 uses large cell-profile datasets to provide finer-grained predictions than broader models, including AlphaGenome.
The Verge - Adam Gayoso said researchers can use it to examine variants found in large populations and investigate their cellular effects.
The Verge - Avsec said newer technologies that capture individual cells and their states at high resolution made it possible to generate the datasets needed to train the model.
The Verge - DeepMind is working to improve its genetics models by generating more data, including data for underrepresented cell states; Avsec said future versions could use data “across multiple species” and potentially contribute to “virtual cells,” digital simulations of biological systems.
(UPDATED November 11, 2026 21:20 UTC)
Google DeepMind unveiled AlphaGenome 2, predicting genetic effects in specific cell types and biological states. It identifies nearly twice as many variants that change gene expression. Researchers can use it to investigate how genetics affects disease.