Researchers from ETH Zurich have developed a new generative artificial intelligence (AI)-based computer process to develop drug molecules based on a protein’s three-dimensional surface.
The new process could revolutionise drug research, making it possible to generate active pharmaceutical ingredients quickly and easily.
The new computer process’ algorithm was developed in collaboration with ETH’s professor Gisbert Schneider and former doctoral student Kenneth Atz, using AI to design new active pharmaceutical ingredients.
Researchers trained the AI model with information from hundreds of thousands of known interactions between chemical molecules and the corresponding three-dimensional protein structures.
The algorithm generates the blueprints for potential drug molecules that can increase or inhibit the activity of proteins with a known three-dimensional shape.
The generative AI then designs molecules that bind specifically to the protein according to the lock-and-key principle to be interacted with.
“This means that when designing a drug molecule, we can be sure that it has as few side effects as possible,” said Atz.
Alongside pharmaceutical company Roche as well as other cooperation partners, the team tested the process by searching for molecules that interact with proteins in the PPAR class – proteins that regulate sugar and fatty acid metabolism.
The AI automatically designed new molecules that also increase the activity of PPARs, reducing the discovery process time. Researchers from Roche then tested these molecules, which showed that the new substances were stable and non-toxic.
“Our work has made the world of proteins accessible for generative AI in drug research,” said Schneider, and is a “real breakthrough for drug discovery”.
Researchers worldwide are currently able to use the algorithm and the software for their own research projects and studies, including a project with the Children’s Hospital Zurich for the treatment of medulloblastomas, the most common malignant brain tumours in children.
