AI uncovers hidden DNA switch that activates most human genes

A group of researchers at the University of California, San Diego have leveraged AI to decode the initiator, a DNA element that marks where a gene’s information begins to be expressed. By measuring gene-expression outcomes for about 500,000 different initiator sequences, they trained a machine-learning system that could pinpoint the characteristic pattern of this switch. Applying the model to the human genome revealed that roughly 60% of genes contain the initiator.

According to professor James T. Kadonaga, the model now offers strong predictions of initiator presence, opening a path to anticipate the impact of harmful mutations. The findings also provide a foundation for engineering synthetic promoters that can precisely control gene activity. Overall, the study highlights the power of combining high-throughput experiments with AI to unravel the complex code governing when and where genes are turned on.