Chalmers researchers have created artificial DNA that controls the protein production of cells using artificial intelligence. The technology can contribute to the development and production of vaccines, drugs for severe diseases, and alternative foods much faster and at a much lower cost than is currently possible.
The genetic code in our DNA is transcribed into the molecule mRNA. It acts as a messenger, telling the cell’s factory which protein and in what quantities it should produce. The process is essential to the operation of cells in all living organisms.
Researchers have worked hard to control gene expression because it can help with the development of protein-based medicines. A recent example is the mRNA vaccine against covid19. It instructs the body’s cells to produce the same protein found on the coronavirus’s surface. The immune system of the body can then learn to form antibodies against the virus. Similarly, one understands the genetic code underlying the production of specific proteins. And then teach the body’s immune system to defeat cancer cells or other complex diseases.
The majority of today’s new drugs are precisely protein-based. But the techniques for producing them are both expensive and slow. Because controlling how DNA is expressed is difficult. Last year, a Chalmers research group led by Alexey Zelezniak, associate professor of systems biology, made an important breakthrough in understanding how much of a protein is made from a specific DNA sequence.
It all started with being able to fully “read” the DNA molecule’s instructions. We have now succeeded in designing our own DNA. It contains precisely the instructions required to control the amount of a specific protein produced. It is the research group’s latest significant breakthrough.
The new method works on the same principle as when an AI generates faces that look like real people. The AI can create completely new realistic faces by learning what a large number of faces look like. The face can then be easily modified by saying, for example. It should look older or have a different hairstyle. However, without AI, programming a convincing face from scratch would have been much more difficult and time-consuming. Similarly, the researchers’ AI has discovered the structure and quantity-regulating code in DNA. The AI then creates synthetic DNA that can be easily altered in the desired direction.
Simply put, the AI receives an order specifying how much of a gene should be expressed. And then “prints” the appropriate DNA sequence.
DNA is a massively long and intricate molecule. It is thus extremely difficult to change it experimentally by reading, changing, reading, and changing again. Years of research are required to find something that works. Then it is far more efficient to train an AI on the principles of navigating DNA. What would otherwise take years can now be completed in weeks or months. This is according to the study’s first author, Jan Zrimec, a former researcher in Aleksej Zelezniak’s group. He is now working at Slovenia’s National Institute of Biology.