AI Decodes 'Initiator' DNA Switch in 60% of Human Genes — SkimNews
Get the Health newsletter
Daily health & science — research, biotech, public health, the studies worth knowing. Free.
- UC San Diego researchers used artificial intelligence to decode the DNA signature of the "initiator," a genetic element that marks where gene expression begins, identifying its characteristic base pattern for the first time.
- Graduate student Torrey Rhyne-Carrigg and Professor James T. Kadonaga's lab measured gene expression across approximately 500,000 different versions of the initiator using high-throughput DNA sequencing, then trained a machine learning model on those results.
- The trained model detected the initiator in roughly 60% of human genes, providing strong predictions of the element's presence or absence.
- Kadonaga said the findings could help researchers anticipate how mutations affecting the initiator alter gene activity and contribute to disorders, and may support the design of tailored synthetic promoters.
- The study was published in the journal Genes (DOI: 10.1101/gad.353623.125).
- Kadonaga framed the work as a step toward decoding the entire human gene expression code across the six billion DNA bases in each cell, calling the initiator model a "small but important part."
Why it matters: The initiator element underlies gene activation across an estimated 60% of human genes, meaning the AI model could give researchers a concrete tool for predicting how specific mutations disrupt gene expression — with stated applications in disease research and synthetic promoter design, as Kadonaga's lab noted. Because the initiator governs when and where genes switch on, decoding its signature moves the field closer to mapping the full gene expression code in human DNA.
Ask SkimNews
