The Mulliken Center for Theoretical Chemistry at the University of Bonn is collaborating with Merck. The three-year collaboration program focuses on developing new tools for computational chemical modeling as well as new representations of molecules.
"We are excited to work with Merck on this project, which will be beneficial for the company and the computational chemistry community in general," said Stefan Grimme, director of the Mulliken Center for Theoretical Chemistry at the University of Bonn. "The close interaction with Merck’s scientists will help us to give the project and the resulting tools the right focus." Grimme, a member of the National Academy of Sciences Leopoldina and an internationally renowned researcher in the specialized field of theoretical chemistry, has developed countless methods and tools with his research group that are now widely used - even beyond the field of computational chemistry.
Merck is leveraging machine learning and artificial intelligence (AI) along all stages of its value chain. Through numerous initiatives and collaborations, the company aims to accelerate the life cycle of its products, break up silos and harness the power of data and digital. "Recent advances have shown the impact that molecular machine learning and AI in general can have in all chemistry-related areas, especially simulation and data-driven drug discovery, materials design and prediction of new formulations," says Jan Gerit Brandenburg, Head of Digital Chemistry at Merck. "With this collaboration, together we want to develop new molecular representations and computational tools that will help us in making drug candidate screenings faster, discover new compounds and predict the performance of materials."
Over the next three years, several PhD students from the Mulliken Center for Theoretical Chemistry will work with the Digital Chemistry team at Merck to identify methods applicable to the company’s entire portfolio of chemicals and pharmaceuticals that would benefit from molecular machine learning techniques. All methods and codes developed within the program will be open source and will thus also benefit the broader scientific community. The program is partly embedded in the German Research Foundation’s (DFG) Priority Programme on Molecular Machine Learning (SPP 2363).