Scientists can now look up what almost any single-letter change in human DNA might do — all 9 billion of them

Your DNA is a book of about 3 billion letters, which means there are more than 9 billion possible single-letter typos — far too many to test one by one in a lab. Google DeepMind has now released the AlphaGenome Atlas, a one-petabyte (a million gigabytes) searchable collection of computer predictions for what each of those tiny changes might do across hundreds of types of human cells — which genes it might switch on or off, and how. Any researcher can explore it free in a web browser, with no code and no charge for non-commercial research. Early users at the Broad Institute used its ranking score to spot an overlooked DNA change behind an unsolved rare-disease case, and a University of Exeter team used it on data from more than 54,000 UK Biobank volunteers to find new links between rare changes and protein levels. The predictions are a map for where to look next, not medical answers — but a good map is exactly what genetics has been missing.

SDG 9 · Industry, innovation and infrastructure — Mechanism: tech innovation