Google DeepMind has launched AlphaGenome Atlas, a free research database containing AI-generated predictions for the effects of all 9 billion possible single-letter mutations in the human genome.
The database is built using DeepMind’s AlphaGenome model and is designed to help researchers better understand both protein-coding DNA and regulatory regions that control how genes are switched on and off.
Predictions Available Instantly
Previously, researchers often had to test genetic variants individually with computational models or study them in laboratories, making large-scale analysis extremely time-consuming.
AlphaGenome Atlas instead provides a precomputed catalogue showing what each possible single-base substitution is expected to do.
DeepMind says analysing all 9 billion possibilities using traditional methods could have taken many human lifetimes.
The company believes the database could make it easier for scientists to study genetic diseases and potentially accelerate research into new treatments.
Building on the Human Genome Project
Pushmeet Kohli, DeepMind’s vice president for research and head of AI for science, said the release gives researchers access to a comprehensive map of human genetic variation simply through a web browser.
He compared the project with the Human Genome Project, which completed the mapping of the human DNA sequence in 2003.
Kohli said scientists had effectively obtained the “book” of the human genome but still faced the challenge of understanding how to read it. AlphaGenome Atlas is intended to help address that problem.
Free for Research Use
AlphaGenome Atlas is available from today for non-commercial research use through a dedicated website.
Google DeepMind plans to offer commercial access through a licensing arrangement with Google Cloud at a later date.
The company has not yet disclosed the pricing or terms of those commercial licences.
DeepMind’s sister company, Isomorphic Labs, which uses AI for drug discovery, will also have access to the database but will require a commercial licence.
Research Paper Released
A paper explaining how AlphaGenome Atlas was developed and how the predictions were generated is also being released through bioRxiv, the preprint repository for biomedical research.
DeepMind says the project is intended to give scientists faster access to information that could otherwise take years of individual experiments to produce.
The post Google’s New AI Atlas Presents Breakthrough for Genetic Disease Research appeared first on ProPakistani.
