Google (GOOGL.US) DeepMind launches AlphaGenome 2: Capable of decoding DNA cells with nearly double the mutation identification capacity

According to information from Zhicheng Finance APP, DeepMind, the artificial intelligence division under Alphabet (GOOGL.US), released its latest AI model, AlphaGenome 2, on Wednesday. This model is designed to decode DNA cells.

DeepMind wrote in a blog post, “By moving from broad tissue averages down to individual cell types and cell states, and by redesigning the model architecture, AlphaGenome 2 has achieved a significant leap in variant prediction performance. Importantly, it enables researchers to predict the effects of variants in disease-related cells that would be missed by broader tissue models and can also predict cell states and disease states previously unseen by the model. Compared to the previous state-of-the-art model, AlphaGenome 2 identifies almost twice as many variants that alter gene expression within cells.”

DeepMind further stated that AlphaGenome 2 has been trained on 68 tissue types and over 100 million single-cell atlases. As a result, it can transform static cellular atlases into genetic prediction engines.

“Surpassing the original model’s reliance on overall tissue averages, AlphaGenome 2 is able to resolve activity within specialized cells—for example, specific cardiac muscle cells or blood immune cells—thus helping identify key DNA variants that only function in certain cell types,” DeepMind added.

This new AI model brings multiple improvements over its predecessor, including predicting gene regulation within specific cells, enhancing the ability to forecast variant effects, and precisely identifying variants associated with diseases. The company cited a recent study revealing that a regulatory switch triggers inflammatory bowel disease by over-activating the ETS2 gene in gut immune cells, a finding that AlphaGenome 2 is capable of replicating computationally.

AlphaGenome 2 is currently available for non-commercial research via the AlphaGenome API, and is offered as a private preview for commercial research use.