Scientists working in the research fields of chemistry and Artificial Intelligence (AI) at the Swiss Federal Institute of Technology in Zurich (ETH) have, together with experts from Roche Pharma Research and Early Development, developed a process that helps to determine the optimal production process for new drugs, according to a meadia release. In this context, they are working at the interface of academic AI research and laboratory automation.
While potential pathways for laboratory synthesis were previously derived from known chemical reactions, with each one tested via experiments as part of a complex trial-and-error process, the new AI model is capable of predicting suitable molecular sites. Borylation was chosen as the chemical activation method for the underlying drug scaffold.
On the subject of borylation, the team led by doctoral student Kennetz Atz and Professor Gisbert Schneider from the Institute of Pharmaceutical Sciences (IPW) conducted a comprehensive worldwide literature search and found 38 particularly trustworthy scientific papers that described a total of 1,380 borylation reactions. This training data set was supplemented by the addition of 1,000 reactions carried out in an automated laboratory operated by Roche’s medicinal chemistry research department.
The AI model generated from this data pool is able to predict the best sites in addition to specifying sites on the scaffolds where activation is not possible. It also provides the optimum conditions for the activation reactions. In the same vein, the team is now keen to find effective models for activation reactions other than borylation. Moreover, Atz is now playing an active part in Roche’s medicinal chemistry research as an AI scientist. “This innovative project is another outstanding example of collaboration between academia and industry and demonstrates the enormous potential of public-private partnerships for Switzerland”, as Gisbert Schneider explains in the press release. ce/mm