Improving empirical force fields: A big-data approach (2018-2021)
Abstract
The ability to simulate the structural and dynamic properties of molecules at an atomic level is at the heart of modern structural biology, computational drug design and materials research. These calculations depend directly on the accuracy with which the forces acting between atoms can be calculated. We will improve the ability of current empirical force fields to reproduce the structural and thermodynamic properties of molecules by exploiting a rapidly expanding in-house database of parameterized molecular structures to develop highly optimized, well-validated, parameters that are both consistent and transferable enabling molecules of any size or complexity to be parameterized with a fidelity currently only possible for simple organics.