Home / Science / AI Decodes DNA's Cell-Specific Secrets
AI Decodes DNA's Cell-Specific Secrets
11 Nov
Summary
- New AI model predicts how DNA mutations affect specific cell types.
- Trained on over 100 million cell profiles across 68 tissues.
- Identifies twice as many gene expression variants as prior models.
Researchers at Google DeepMind have introduced AlphaGenome 2, an advanced AI model designed to predict the effects of DNA mutations on specific cell types and biological states. This successor to AlphaGenome offers a significantly more detailed perspective on genetic variations by analyzing how genes are switched on and off within different cells.
The model was trained using data from over 100 million individual cell profiles spanning 68 tissues. This extensive training allows AlphaGenome 2 to transform static cell atlases into a dynamic predictive engine for human genetics, identifying nearly double the number of variants that alter gene expression compared to its predecessor.
Notably, AlphaGenome 2 can make predictions for new cell types and biological states without requiring retraining. It also assists researchers by identifying gaps in experimental data, guiding future laboratory investigations. This capability is crucial for tying human genetics to cellular biology, enabling scientists to pinpoint the cellular-level effects of genetic variants linked to diseases.
Google DeepMind is making AlphaGenome 2 available for non-commercial research via its API, with a limited private preview for commercial use. This development represents a significant step in using AI for scientific and medical advancements, with potential future applications including the creation of virtual cells.
