Google DeepMind unveils AlphaGenome 2 for disease

Google DeepMind has released AlphaGenome 2, an AI model built to predict how changes in DNA affect different cell types. The company says the system offers a more detailed view of gene regulation and disease mechanisms, which could help researchers identify new treatment targets. The work fits a wider effort to apply foundation-model techniques to biology and genomics, where large models can learn patterns across vast amounts of genetic and cellular data. The central challenge is linking genetic variation to cellular function at scale. Many disease-associated mutations lie outside protein-coding regions, so understanding their effects requires better predictions of how DNA controls gene activity across tissues and cell states. If AlphaGenome 2 performs well in independent validation, it could become a useful tool for drug discovery, functional genomics, and precision medicine. Researchers would still need experimental confirmation, and the model's limitations will matter, including how well it generalizes across populations and rare cell types. Still, the release underscores how AI is increasingly being used not just to analyze biological data but to generate hypotheses that guide lab work. For DeepMind, it is another step in positioning AI as a research partner in the life sciences.