AI Decodes Human DNA 'On Switch' for Gene Activation

Artificial intelligence (AI) has decoded a key DNA sequence that acts as an "on switch" for human genes. This breakthrough brings scientists closer to understanding how the human genome controls gene activity. The research focused on an important DNA element called the "initiator."

Researchers at the University of California San Diego used high-throughput DNA sequencing. They measured gene expression across approximately 500,000 different versions of the initiator. These results trained a machine learning system to identify the characteristic DNA pattern of the initiator.

The AI model identified the initiator in about 60% of human genes. This is the first time AI models have provided strong predictions for the presence or absence of the initiator in human genes. The findings were published in the journal *Genes*.

This discovery could help predict how mutations affecting the initiator might alter gene activity. Such alterations can contribute to various disorders, including cancer. The study's data and AI models may also aid in designing synthetic promoters. These sequences can switch genes on or off for specific purposes.

This work represents a step forward in combining laboratory experiments and AI to decipher information embedded in human DNA. The ultimate goal is to create an AI model for the entire gene expression code. This would allow predictions of gene activity across different individuals and genetic variants.