Transferable Machine Learning Interatomic Potential for Bond Dissociation Energy Prediction of Drug-like Molecules

Add to your list(s) Download to your calendar using vCal

• Elena Gelzinyte, University of Cambridge
• Monday 12 June 2023, 14:00-14:30
• Zoom link: https://zoom.us/j/92447982065?pwd=RkhaYkM5VTZPZ3pYSHptUXlRSkppQT09.

If you have a question about this talk, please contact Dr Venkat Kapil.

We present a transferable MACE interatomic potential that is applicable for open- and closed-shell drug-like molecules containing C, H, O chemical elements. We explore MACE transferability to the COMP6 dataset and show that it reaches accuracy on par with state-of-the-art transferable ANI2x potential for closed shell molecules. An accurate description of radical species extends the scope of possible applications to reaction energy prediction, for example in the context of Cytochrome P450 (CYP) metabolism. MACE potential reaches similar accuracy on two CYP substrate datasets, with open-and closed-shell structures relevant to CYP metabolism. We apply MACE to aliphatic C-H bond dissociation energy prediction where it reaches RMSE below 1.6 kcal/mol and a better rank prediction than currently used AM1 semi-empirical method.

This talk is part of the Lennard-Jones Centre series.

This talk is included in these lists:

• Hanchen DaDaDash
• Lennard-Jones Centre
• Zoom link: https://zoom.us/j/92447982065?pwd=RkhaYkM5VTZPZ3pYSHptUXlRSkppQT09

Note that ex-directory lists are not shown.