Researchers have created an artificial intelligence system that can predict the ways in which drug molecules can undergo chemical changes. A collaborative team from Germany's Technical University of Munich (LMU), ETH Zurich (ETH Zurich) and Basel-based Roche Pharmaceutical Research and Early Development (pRED) has used artificial intelligence (AI) to design a new technology for predicting the best way to synthesize drug molecules.
"This approach has the potential to significantly reduce the number of laboratory experiments required, thereby increasing the efficiency and sustainability of chemical synthesis," said David Nippa, first author of the corresponding paper, which has been published in the journal Nature Chemistry. Nippa is a PhD student in the Department of Chemistry and Pharmacy at LMU and in the research group of Dr. David Konrad at Roche.
Active pharmaceutical ingredients usually consist of a framework to which functional groups are attached. These groups have specific biological functions. To achieve new or better medical effects, functional groups need to be changed and added to new positions in the framework. However, this process is particularly challenging in the field of chemistry because the framework itself, which is composed mainly of carbon and hydrogen atoms, is almost inactive.
One method of activating frameworks is the so-called borylation reaction. In this process, chemical groups containing boron are attached to the carbon atoms of the framework. This boron group can then be replaced with various groups with medical effects. Although boronation reactions have great potential, they are difficult to control in the laboratory.
Together with Kenneth Atz, a doctoral student at ETH Zurich, David Nipa developed an artificial intelligence model that was trained on trusted scientific works and experimental data from Roche's automation laboratories. It can successfully predict the boration position of any molecule and provide optimal conditions for chemical transformations. "Interestingly, the predictions improved when three-dimensional information about the starting materials was taken into account, rather than just their two-dimensional chemical formulas," Artz said.
This method has been successfully used to identify positions in existing active ingredients where additional reactive groups can be introduced. This helps researchers more quickly develop new, more effective variants of known drug active ingredients.
Compiled from /scitechdaily