25 November 2022

Proteins to order

Artificial Intelligence Adapts synthetic DNA for drug development

Tatiana Matveeva, "Scientific Russia"

With the help of artificial intelligence, Swedish scientists have developed synthetic DNA that controls the production of protein by cells. The new method may make it much faster and cheaper in the future to create and produce medicines for serious diseases, as well as alternative dietary proteins, according to Chalmers University of Technology. The results of the work are presented in the journal Nature Communications (Zrimec et al., Controlling gene expression with deep generative design of regulatory DNA).

The way our genes are expressed is a fundamental process for the functioning of the cells of all living organisms. Briefly, the mechanism of expression can be described as follows: the genetic code in DNA is transcribed into a matrix RNA (mRNA) molecule, which tells the cellular "factory" which protein to produce and in what quantities.

Gene expression can be controlled to some extent, that is, to set precise instructions for DNA (without affecting the genetic code itself) so that the required amount of a certain protein is produced. With the help of special preparations, you can tell the cells which proteins to produce. This is how, for example, the mRNA vaccine against Covid-19 works, which tells the body's cells to produce the same protein as on the surface of the coronavirus. Then the immune system can learn to produce antibodies against the virus. Similarly, you can teach the body's immune system to defeat cancer cells or other complex diseases if you understand the code underlying the production of certain proteins.

The production of such protein drugs today is a long, complex and expensive process, according to researchers from Chalmers University of Technology. They propose to simplify and speed it up – with the help of artificial intelligence. So, the team trained the algorithm on a variety of examples of DNA molecules – with different structures and regulatory code. Having learned, the AI has designed synthetic DNA, which can be easily corrected on request and thereby affect gene expression. In other words, the AI reported which part of the gene was needed, and then it "printed" the corresponding DNA sequence.

"Manually" – experimentally – it is not easy to make changes to a long DNA molecule. It is much more effective to let AI learn the principles of DNA navigation. What used to take years, can now be done in weeks or even in a matter of days, says lead author Jan Zrimec, a researcher at the Chalmers Institute of Technology and the National Institute of Biology of Slovenia.

The researchers developed their method on yeast Saccharomyces cerevisiae, whose cells resemble mammalian cells. In future studies, scientists plan to test the method on human cells.

Portal "Eternal youth" http://vechnayamolodost.ru

Found a typo? Select it and press ctrl + enter Print version