As computer simulations are increasingly used to complement or replace experiments, highly accurate descriptions of microscopic interactions are required to achieve realistic predictions. The questions of how to objectively measure model quality and how to relate the form and parameters of a force field to experimental data are therefore of prime interest. To address these issues, we use the concept of statistical distance to define a measure of dissimilarity between a pair of statistical mechanical systems, i.e., a model and its target, and show that its minimization leads to general convergence of the model’s measurable properties to those of the target.1 We exploit this feature to define a rigorous basis for the development of accurate and robust effective molecular force fields that are inherently compatible with coarse-grained experimental data. The new model optimization principles and their efficient implementation are illustrated through selected examples, which demonstrate the higher robustness and predictive accuracy of the approach compared to other currently used methods, such as force matching2 or relative entropy minimization.3 1. Vlcek, L.; Chialvo, A. A., Rigorous force field optimization principles based on statistical distance minimization. The Journal of Chemical Physics 2015, 143 (14), 144110. 2. Ercolessi, F.; Adams, J. B., Interatomic potentials from 1st principles calculations-the force-matching method. Europhys. Lett. 1994, 26 (8), 583-588. 3. Shell, M. S., Systematic coarse-graining of potential energy landscapes and dynamics in liquids. J. Chem. Phys. 2012, 137 (8).

Statistical distance between thermodynamic systems as a theoretical basis for force field development

251st National Meeting (2016)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.