Google DeepMind unveiled AlphaGenome 2, an AI model built to predict how changes in DNA play out in different kinds of cells.
The model succeeds AlphaGenome, DeepMind's earlier genome-predicting system. The new version targets a basic complication in biology: nearly every cell in the body carries the same DNA, yet heart muscle cells, skin cells and immune cells fighting an infection each switch different genes on and off. Researchers at DeepMind designed AlphaGenome 2 to give scientists a more detailed view of those cell-by-cell differences.
That matters because a DNA change rarely means anything on its own. Whether it does harm depends on which cells read it and which genes those cells have active. A model that works at the level of individual cell types gets closer to the question researchers care about, which is where in the body a genetic change goes wrong. DeepMind's stated hope is that this detail could offer clues to the causes of disease and point toward new treatments.
There is a gap between prediction and proof, though. A model that forecasts which genes a cell will switch on is a research tool, and a flagged variant still needs lab work before anyone calls it a cause of disease. AlphaGenome 2 narrows where to look. It does not replace looking.