Researchers in Sweden have succeeded in creating artificial DNA that can control the cellular production of proteins with the aid of artificial intelligence (AI). This, they say, could aid future drug development.
Artificial Intelligence
The new technology could speed up the development of a variety of vaccines and drugs for serious medical conditions. It would make the development of alternative food proteins more rapid and less costly, according to researchers.
Read also: Artificial Intelligence Accurately Diagnoses Brain Tumors in Austrian Study
This AI-designed DNA was reported in a paper published in Nature Communications. The authors were led by researchers from Sweden’s Chalmers University of Technology.
Regulating gene expression
The functionality of all living organisms’ cells depends on how genes are expressed. There is a transcription of the DNA’s genetic code to the messenger RNA (mRNA), which determines the protein to produce as well as the quantity.
There has been a lot of focus among researchers on controlling gene expression. This is due to its importance in the development of protein-based drugs.
The Covid-19 mRNA vaccine is a recent example of one based on proteins. This instructs cells in the body to make protein similar to that present on the coronavirus’s surface. Through this, the immune system learns enough to make antibodies against the virus.
In the same way, the immune system can be taught to fight cancer cells and other diseases that are usually difficult to treat once researchers can grasp the underlying genetic code for the production of certain proteins.
The larger proportion of drugs that became available more recently is based on proteins. Their productions are, however, usually long-drawn-out and expensive due to the difficulty in controlling DNA expression.
“DNA is an incredibly long and complex molecule,” explains Jan Zrimec, the study’s first author. “It is thus experimentally extremely challenging to make changes to it by iteratively reading and changing it, then reading and changing it again. This way it takes years of research to find something that works.”
The production of protein-based drugs or alternative food proteins can become so expensive that getting a return on investment is practically impossible.
Chalmers researchers led by Aleksej Zelezniak had a breakthrough in their understanding and regulation of the amount of protein produced from a specific DNA sequence last year.
A more efficient approach
Researchers say their new method, which was developed in Saccharomyces cerevisiae with cells similar to those in mammals, makes things faster. It takes just weeks or even days to learn what would have otherwise taken years, according to Zrimec.
Read also: How Artificial Intelligence Can Improve Clinical Research
The underlying principle is similar to that involved when generating human-like faces with AI. In the latter case, the system studies a variety of faces to learn what they look like and then uses the knowledge to create faces that look like real humans. This makes creating natural-looking faces a lot easier.
In the current work, researchers taught their AI the structure of DNA as well as its regulatory code for it to design synthetic DNA that can be modified for specific gene expression. Put differently, the AI was instructed on how much of a gene is required for it to “print” the right DNA sequence.
“First it was about being able to fully ‘read’ the DNA molecule’s instructions,” said Zelezniak, an associate professor of Systems Biology at Chalmers. “Now we have succeeded in designing our own DNA that contains the exact instructions to control the quantity of a specific protein.
Read also: An Artificial Intelligence Is Now Able to Predict Cardiovascular Diseases Through The Eyes
The researchers said their technology makes for a lot more efficient production of proteins that could be marketed.
Next, the team wants to use this AI on human cells. The researchers hope their work will prove helpful to the development of both new and existing drugs.
References
Controlling gene expression with deep generative design of regulatory DNA
FEEDBACK:
