Google DeepMind unveils AlphaGenome 2
Google DeepMind has announced AlphaGenome 2, a sequel to its genomics model that aims to predict how DNA shapes cell behaviour and disease, though the material we received includes no performance details.
Google DeepMind has announced AlphaGenome 2, describing it as a model for "predicting how DNA shapes cells and disease." The post's URL describes the work as deciphering how cells read DNA to better understand disease.
This continues DeepMind's genomics work. As we reported earlier, the lab has mapped 9 billion human DNA variants.
The text available to us contains only the post's acknowledgements, not its technical content. We have no details on architecture, training data, benchmark results, how it compares with the original AlphaGenome, or whether and how researchers can access it.
The acknowledgements list 24 core contributors. They include DeepMind CEO Demis Hassabis, Pushmeet Kohli and Žiga Avsec, along with Adam Gayoso, Natasha Latysheva and Jun Cheng. The team also thanks a second group for "support and feedback," including Sarah Teichmann, Leopold Parts, Junyue Cao and Joseph Powell. The material does not say how those people were involved or where they work.
We will update this story when the model's claimed capabilities and evaluation results can be checked against the full announcement.