Google DeepMind unveils AlphaGenome 2 AI model to predict DNA variant impacts at single-cell resolution

Executive Briefing
“On November 11, 2026, Google DeepMind introduced AlphaGenome 2 to predict DNA variant...”
Key Takeaways
- 1.Google DeepMind launched AlphaGenome 2 on November 11, 2026, shifting predictions from broad tissues down to individual cell types and states.
- 2.The model was trained on over 107 million single-cell profiles across 68 human and mouse tissues, incorporating DNA methylation data.
- 3.On the TenK10K dataset, the model identified 30% of pathogenic variants compared to 16% for the fine-tuned first version.
On November 11, 2026, Google DeepMind introduced AlphaGenome 2 to predict DNA variant impacts at single-cell resolution instead of averaging across tissues. The model enhances disease variant identification through an advanced conditional architecture.
Key Developments and Technical Context
Google DeepMind launched AlphaGenome 2 on November 11, 2026, shifting predictions from broad tissues down to individual cell types and states.
The model was trained on over 107 million single-cell profiles across 68 human and mouse tissues, incorporating DNA methylation data.
On the TenK10K dataset, the model identified 30% of pathogenic variants compared to 16% for the fine-tuned first version.
Strategic Analysis and Industry Implications
The conditional architecture prevents the dilution of computational capacity, allowing the system to capture subtle differences between closely related cell types.
Providing the tool for free via an API helps domestic research groups access advanced biotechnology without investing in heavy server infrastructure.
Outlook and Key Factors to Watch
AlphaGenome 2 has not yet been validated or approved for practical clinical diagnostic use.







