Google DeepMind released a free tool that helps scientists find answers for rare diseases in children faster. It’s called AlphaGenome Atlas, and it came out on September 8, 2026. Researchers open a simple website, with no coding required, and the tool sorts a huge list of known issues with DNA so the ones most likely to matter show up first.
Here’s what that means. Some kids have seizures, and no genetic test explains why. Parents leave the hospital with a long list of possible issues with the child’s DNA, and almost no idea which ones matter.
DNA carries the body’s basic instructions, written in four letters. A known issue with DNA, in this story, is often one letter sitting in the wrong place. Most of those letter changes do little or nothing. A few change how the body works. Finding the important ones used to take more time and lab work than most teams had.
AlphaGenome Atlas ranks about 9 billion possible single-letter DNA changes, and it guesses what each one might do inside the cell. DeepMind also made a simple score to help with the ranking. It blends two of their AI models into one number, so a scientist spends time on the changes most worth testing. Academic labs get free access now. A paid option for companies is planned later.
And people are already using it. Scientists at the Broad Institute tried it on rare diseases older tests had left unsolved. DeepMind’s blog says they found a missed DNA change in a gene tied to a serious seizure condition. The AI predicted how the change altered the body’s message. Lab tests agreed. For a family waiting on a name for their child’s illness, a short ranked list with a clear idea to test is better than walking out with nothing.
Tools like this point toward real progress in everyday life. More families leave the clinic with a clearer name for what is going on, and they get there sooner. Doctors spend less time guessing which DNA change to test next. Drug and treatment research moves faster when the starting list is shorter and better ranked. None of that happens overnight, and lab work still has to confirm what the AI suggests. The direction still matters, with faster answers for rare disease, better starting points for the science that follows, and more chances for care to catch up to the kids who need it.