Propagation of uncertainty in physicochemical data to force field predictions

Abstract

The solvation free energy (SFE) is a key property in the thermodynamics of chemical processes. It can be evaluated using molecular simulations with good statistical accuracy. However, force field predictions exhibit systematic errors due to uncertainties in the parametrization. Here we evaluate how the uncertainty in physicochemical data underlying force fields propagates to SFE predictions. We find that the data contribution to the uncertainty in SFE is up to 25 times larger than the statistical uncertainty. The total uncertainty in the SFE in water is higher than in cyclohexane.