AI deciphers the 'on switch' in human DNA
Machine learning model reveals the DNA sequence pattern controlling gene expression.
1Core mechanism of gene activation
Healthy growth requires thousands of genes activating at the right time and place. Special DNA regions coordinate this process, guiding production of enzymes and proteins. Errors in gene activation can lead to cellular dysfunction and diseases like cancer. UC San Diego researchers focused on the 'initiator' DNA element that plays a key role in this process.
2500,000 initiator variants analyzed
Using high-throughput DNA sequencing, researchers measured gene expression across approximately 500,000 initiator versions. The team, led by graduate student Torrey Rhyne-Carrigg, used this data to train a machine learning model. The AI successfully identified the characteristic DNA pattern of initiators. Scans revealed this sequence exists in about 60% of human genes.
3AI model makes first accurate predictions
Professor James Kadonaga from UC San Diego reported that their AI models can predict initiator presence in human genes with unprecedented accuracy. The model decoded the DNA base sequence pattern of initiators. This breakthrough will help understand how mutations affecting initiators alter gene activity. It could also advance synthetic promoter design for controlling genes.
4New insights into genetic diseases
The study shows how initiator mutations can disrupt gene expression and cause various disorders. AI models will enable predicting effects of such mutations. For instance, we'll better understand genetic mechanisms underlying complex diseases like cancer. The research may also contribute to developing personalized gene therapies for various conditions.
5Potential for synthetic gene control systems
Another significant outcome is supporting design of synthetic promoters that can switch genes on/off. This technology could enable customized genetic control systems for specific functions. Examples include circuits regulating insulin production in diabetics or inhibiting cancer cell growth. Such advances could revolutionize personalized medicine.
6Power of combining lab work with AI
The research demonstrates how combining laboratory experiments with artificial intelligence can reveal hidden information in DNA. Kadonaga notes this approach could help decode the gene expression pattern hidden in human DNA's 6 billion base pairs. The initiator AI model represents a small but crucial piece of this larger puzzle.
7Goal to decode complete gene expression
Kadonaga explains their initiator AI model only deciphers a small part of the gene expression code, but expects to expand these models in the near future. A complete gene expression AI model would predict activity of genetic variants across different individuals. This could enable breakthroughs in personalized medicine and treatments.
8New era in genetic research
This study highlights the potential of AI-assisted genetic research. Understanding DNA's complex regulatory mechanisms could revolutionize both basic science and clinical applications. The work also emphasizes interdisciplinary collaboration - new discoveries emerge at the intersection of molecular biology, computer science and data analysis.