Google DeepMind released AlphaGenome Atlas, an approximately one-petabyte dataset containing predictions for more than 9 billion single-letter human-genome changes.

Google DeepMind released AlphaGenome Atlas, which HPCwire describes as an approximately one-petabyte dataset. It contains predictions for more than 9 billion possible single-letter changes in the human genome.

The source frames the resource around studying how changes in DNA affect the way the body works. The count refers to possible genomic changes covered by predictions—not billions of experimentally validated findings, individuals or clinical diagnoses.

The petabyte figure describes the dataset’s size, not the compute used to generate it. The supplied excerpt does not disclose GPU requirements, inference costs, prediction-accuracy measurements or experimental-validation coverage. It establishes the release and scale of a predictive research dataset, rather than a verified medical outcome.