Google's new AI predicts how DNA changes affect specific kinds of cells

Google DeepMind has released AlphaGenome 2, an AI model that predicts how small changes in DNA affect specific kinds of human cells, such as a heart muscle cell or an immune cell. It finds nearly twice as many DNA variants that change gene activity as the company's earlier model. It is a research tool, not a medical test.

Why cell type matters

Every cell in the body carries the same DNA. But each cell type switches different genes on or off to do its job. A tiny DNA change might matter in one cell type and do nothing in another. Predicting those effects is still an unsolved problem, and it is a key driver of human traits and disease.

What changed

The first AlphaGenome, launched last year, predicted variant effects across broad tissues. Its companion platform, AlphaGenome Atlas, holds predictions for all 9 billion possible single-letter changes in the human genome. Broad tissue averages can hide what happens in one specialized cell type.

AlphaGenome 2 zooms in. It looks at one condition at a time, meaning a tissue, a cell type or a cell state, alongside a DNA sequence a million letters long. It was trained on more than 100 million individual cell profiles across 68 tissues, using human and mouse data. It also adds a readout called DNA methylation, which researchers had widely requested.

New cells without retraining

Scientists can upload a snapshot of gene activity from just a few dozen cells. The model can then predict how DNA variants act in that new cell state without being retrained. Google says it can match or beat models trained directly on those datasets. One example is a COVID-19 mRNA vaccine study, where it was tested on cell data it had not seen before.

The model can also fill in missing data. If a scientist supplies one type of measurement, it can predict others, such as how open the DNA is for reading. That could help researchers choose which lab experiments to run next.

Who can use it

• Non-commercial researchers can use it now through the AlphaGenome API.
• Commercial research is in private preview.
• It is aimed at scientists, not patients or doctors.

What is not known yet

The material describes better predictions, not new treatments. It does not say which diseases the tool has helped explain or when any medical benefit might arrive. Any real discovery would still need testing in the lab. For most people, the effect will be indirect and gradual, through faster research into the causes of disease.