ETH researches artificial intelligence for protein synthesis for drugs

Researchers from the Department of Chemistry and Applied Biosciences at the Swiss Federal Institute of Technology in Zurich (ETH) have developed an algorithm that can use artificial intelligence (AI) to visualize new active pharmaceutical ingredients. According to a press release, this involves developing appropriate active substances for three-dimensional models of any protein, which can increase or decrease the activity of the protein as a drug. In the model, Professor Gisbert Schneider's group used a generative AI that was trained with information on hundreds of thousands of known structures of protein molecules, including their interactions with other chemical structures. The aim is to generate active substances that produce interactions at a desired site but not at others. The aim is to rule out side effects of medication as far as possible.

The first practical tests, which were carried out together with the pharmaceutical company Roche, were successful. This enabled the design of molecules that interact with representatives of the PPAR protein class - proteins that regulate sugar and fat metabolism. Tests showed that the generated molecules increased the activities of PPAR in the same way as previously commercially available drugs against diabetes mellitus. Further clinical trials are planned, for example at the Zurich Children's Hospital for the treatment of medulloblastomas, the most common malignant brain tumors in children.

"With our work, we have made the world of proteins accessible for generative AI in drug discovery," Gisbert Schneider is quoted as saying in the press release. "The new algorithm has enormous potential. It is interesting for all medically relevant proteins in the human body for which no chemical compounds are yet known that interact with them." ce/eb