
16%→30%Variant detection rate vs previous model · Google DeepMind’s own evaluation
AlphaGenome 2 predicts DNA variant effects in specific cells
AlphaGenome 2, Google DeepMind’s DNA model, identified 30% of gene-activity-altering variants versus 16% for its single-cell-tuned predecessor in company tests on TenK10K, an immune-cell dataset. It is available through the AlphaGenome API for non-commercial research, with commercial research access in private preview.
Sources
Official releaseAlphaGenome 2: Deciphering how cells read DNA to better understand disease — Google DeepMindWritten by AI from the sources