About the news
Trained on more than 100 million individual cell profiles across 68 tissues, it identifies nearly twice as many causal variants that influence gene expression compared with the prior model (e.g., 30% vs 16% on TenK10K immune-cell benchmark at high precision).
Availability
It supports expression conditioning on a user’s own single-cell data without retraining and is available via the AlphaGenome API for non-commercial research (commercial preview). The model is positioned as a research tool to accelerate understanding of disease mechanisms at the cellular level.



