AI Model Predicts Gene Activity Across Human Cell Types

Researchers develop a groundbreaking AI system capable of uncovering cellular mechanisms and advancing disease research.

The AI model, trained on over 1.3 million cells, predicts gene expression with high accuracy in previously unseen human cell types.
The system has identified how specific mutations drive inherited pediatric leukemia, confirmed through laboratory experiments.
Researchers aim to use the model to explore the genome's 'dark matter,' where most cancer-related mutations occur, offering new insights into disease mechanisms.
The approach models gene regulation similarly to how language models like ChatGPT learn grammar, enabling predictions of cellular behavior.
This innovation marks a shift toward predictive biology, opening avenues for understanding diseases and identifying potential treatment targets.