Overview
- AlphaGenome uses a transformer architecture trained on public experimental data to forecast how small DNA changes alter gene activity and molecular processes.
- The model went live on June 25, 2025 for noncommercial use and DeepMind plans to publish full technical details in the coming months.
- Researchers can use AlphaGenome to prioritize genetic variants in silico, reducing the need for time-consuming lab experiments.
- Experts say the tool could accelerate diagnosis of rare genetic disorders and help identify cancer-driving mutations by modeling their impact on gene expression.
- DeepMind describes AlphaGenome as a milestone toward a fully virtual laboratory for drug discovery and genome design.