UCSD researchers use AI technology to discover new DNA promoter codes

Recently, researchers from the University of California, San Diego (UCSD) published a study in the journal Nature, introducing their use of artificial intelligence technology to discover a new DNA promoter code, the downstream core promoter region DPR, which is expected to be used in biotechnology and biomedical applications to control gene initiation.

According to researchers, DNA sends precise instructions after receiving coding guidance from the four bases A, C, G, and T, and human genes are activated to take action upon receiving the instructions. Currently, it is known that nearly 25% of genes are transcribed by TATAA-like sequences (i.e. "TATA boxes"), but due to the high possibility of DNA base sequences, how the remaining 75% of genes are activated has always been a mystery.

Researchers evaluated the DPR activity of 500000 random DNA sequences and selected 200000 of them to generate a machine learning model that can accurately predict DPR activity in human DNA. In addition, a similar machine learning model that can recognize TATA box sequences was also developed. Researchers evaluated thousands of known TATA boxes and DPR activity test samples using the new model and found that its predictive ability was "excellent". The research results show that there is indeed DPR activity in human genes, and its incidence is comparable to that of TATA boxes. According to the research author and distinguished professor of the Department of Biological Sciences at UCSD, Kadokawa, DPR can activate approximately 1/4 to 1/3 of genes. In addition, researchers have also found a fascinating duality between DPR and TATA boxes, where genes initiated by TATA box sequences lack DPR sequences, and vice versa.

Kadokawa stated that the reason why DPR was not discovered before was because it did not have obvious sequence patterns. The encrypted information hidden in DNA sequences gives it DPR properties, which humans cannot decipher, but machine learning models can. Using artificial intelligence to analyze DNA sequence patterns should enhance researchers' understanding and control of gene initiation in human cells.

Prev: What are the benefits of a facial recognition visitor system?