UC San Diego researchers decode hidden DNA initiator sequence using AI

AI digestMimile.ai EditorialAugust 23, 2026 at 05:00 AM

This digest was compiled by AI from multiple sources — links to the originals are below.

UC San Diego researchers decode hidden DNA initiator sequence using AI

Researchers at UC San Diego have decoded the DNA sequence pattern of the initiator, a key gene expression element, using machine learning. The AI model predicts the initiator's presence in roughly 60% of human genes. The findings could help anticipate how mutations affect gene activity and disease.

Key Facts

The study measured gene expression activity across approximately 500,000 different versions of the initiator DNA element.
The AI model found that roughly 60% of human genes contain the initiator sequence.
The research was led by graduate student researcher Torrey Rhyne-Carrigg in the laboratory of Professor James T. Kadonaga at UC San Diego.
The findings may support the design of synthetic promoters with tailored functions for switching genes on or off.

AI Decoding Method

The team used high-throughput DNA sequencing to measure gene expression activity across approximately 500,000 different versions of the initiator. Those results trained a machine learning system to identify the characteristic DNA pattern associated with the initiator. Once the model decoded that signature, the team searched human genes for the sequence and found that roughly 60% contain the initiator. Professor James T. Kadonaga stated that the AI models provided, for the first time, strong predictions of the presence or absence of the initiator in human genes.

Potential Applications

The findings could help researchers anticipate how mutations affecting the initiator may alter gene activity and contribute to a range of disorders. The study's data and AI models may also support the design of synthetic promoters, sequences that can switch genes on or off, with functions tailored for specific purposes. Kadonaga noted that the work is a step forward in the combined use of laboratory experiments and AI to decipher information embedded in human DNA. He expressed optimism about expanding AI models of the human gene expression code in the not-too-distant future.

1 source

UC San Diego researchers decode hidden DNA initiator sequence using AI

sciencedaily.com

ScienceDaily

Science

AI model predicts DNA binding with improved accuracy

Researchers developed an AI model that predicts which DNA sequences bind to each other. The model improves understanding of hypercomplex binding relationships. This has applications in diagnostics and DNA computing.

Science

DNA repair enzymes favor specific sequences, shaping human genome mutation patterns

DNA repair enzymes’ sequence preferences influence where mutations accumulate in the human genome, a Weizmann Institute study finds. Published in Nature Communications, the research shows that even genetic letters five positions away from a damage site affect enzyme binding, explaining regional mutation rate differences.

Health

Genome-scale stem cell map reveals gene function guide

Bioengineers at the University of California San Diego have created a genome-scale reference map detailing how individual genes control human stem cell functions. The open-access resource covers 2.5 million stem cells and could enable virtual cell models for complex diseases.

Science

University of Vienna maps 600-million-year 'evolutionary highways' in animal genomes

University of Vienna researchers compared more than 5,800 chromosome-scale genomes from 4,454 species across 19 animal phyla. The study, published in Science Advances, found that animal genome evolution follows restricted, irreversible pathways the team calls 'evolutionary highways.' The findings provide a framework for understanding and conserving animal biodiversity.

UC San Diego researchers decode hidden DNA initiator sequence using AI

Key Facts

AI Decoding Method

Potential Applications

Read more

Read more