AI Reveals Hidden Code Behind Gene Activation

A new artificial intelligence model has uncovered part of the hidden code that controls human gene activation.

Researchers say precise activation of tens of thousands of genes is essential for healthy growth and development. Specialized regions of DNA help coordinate the genetic sequences that produce enzymes, hormones, proteins and other vital components that support cell structure and function. When these genes do not activate properly, cells may malfunction and contribute to disorders, including cancer.

To better understand the DNA sequences that make gene activation possible, scientists at the University of California, San Diego focused on a key part of DNA known as the promoter. This region contains the instructions where coded genetic information is first translated into functional products.

In the new study, the researchers used high-throughput DNA sequencing to measure gene expression activity across nearly 500,000 different promoter variants. They then applied machine learning to build an AI model capable of decoding promoter DNA patterns. Once the promoter identity was revealed, the team was able to search for recognizable sequences and found that about 60 percent of human genes contain promoters.

The discovery could help researchers predict the effects of DNA mutations linked to promoter-related disorders. The data and models from the study may also be used to design synthetic promoters, DNA sequences that switch genes on and off with customized functions.

The work marks another step forward in combining experiments and artificial intelligence to decode the information stored in human DNA. With a complete AI model of the gene expression code, scientists could eventually predict gene activity across different individuals. The new promoter model covers only one small but important part of that system, but researchers say it may help pave the way for broader AI models built on the human gene expression code.

Source: ISNA