Scientists face a big challenge with the human genome. They can read its letters, but understanding their roles is much harder. Only about 2 percent of the genome codes for proteins. The rest acts like a huge control panel, deciding when and where genes turn on. With about 9 billion possible single-letter changes, testing them all in a lab just isn’t possible.

So, Google DeepMind asked a new question: what if we could predict what those changes might do?

This question led to the AlphaGenome Atlas (https://deepmind.google.com/science/alphagenome/atlas?), a detailed map of nearly every possible single-letter change in the human genome. Instead of checking each change one by one, researchers can use the Atlas to spot the ones most likely to matter. The AlphaGenome Variant Impact (AVI) score works like a trail marker, pointing scientists toward the changes worth a closer look.

This is where the Atlas gets especially useful. Much of the genome lies outside the protein-coding regions, where DNA acts as a switch or controller for genes. AlphaGenome lets researchers explore these areas and see how small changes could affect gene activity.

An atlas isn't the destination; it's a guide.

The AlphaGenome Atlas doesn’t replace experiments or solve every mystery. Instead, it helps scientists decide where to begin. From billions of possibilities, it turns the vast genetic landscape into something researchers can start to explore.

There are nine billion possibilities, a vast map, and maybe among them therWith nine billion possibilities and a huge map to explore, there may be clues hidden here to some of medicine’s toughest mysteries.

More at https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/alphagenome-atlas-a-predictive-map-of-every-possible-dna-letter-change-in-the-human-genome/alphagenome-atlas.pdf