Researchers at UC San Diego have used artificial intelligence (AI) to study a short stretch of DNA called the initiator. The initiator is the place where a gene begins the process of being turned into a working product such as a protein. Correct control of this process is essential for normal cell function. When it goes wrong, problems including cancer can arise.
To find the exact DNA sequence pattern of the initiator, scientists measured the activity of about five hundred thousand slightly different versions of this region. They then used machine learning to find patterns and build a model that could recognize the true initiator signature. The model showed that roughly sixty percent of human genes contain this sequence.
How the discovery can be used
With the initiator sequence now known, it becomes possible to search human DNA for mutations that might disrupt gene activation and contribute to disease. The same information can also help design synthetic promoters, artificial DNA sequences that turn genes on or off in controlled ways. The researchers note that this is an early step toward understanding the larger gene expression code written in the six billion DNA bases of each human cell. That code determines when, where, and how strongly every gene is activated. An AI model of the full code could one day predict how gene variants behave in different people.
The work combines careful laboratory experiments with computational analysis. High-throughput DNA sequencing generated the large data set, and the machine learning model translated those measurements into a clear sequence pattern that had not been fully defined before. The study provides both a practical tool for examining disease-related mutations and a demonstration that AI can help decode important regulatory signals hidden in human DNA.
The researchers have described the methods and results of this study in a paper published in Genes and Development.