A New Semi-Explicit Solvation Model: Fast Physics for Better Results

Abstract

Computational physicists, chemists, and biologists have a critical need for better models of water and aqueous solutions. We present an exciting new solvation model called Semi-Explicit Assembly, which combines the speed of the fastest continuum models available with the strong physical basis and discrete water treatment afforded by explicit solvent simulations. We base our model on several simple physical properties of water as a solvent, collected directly from explicit solvent simulations for individual atomic solutes. As a first test and application of our method, we compute solvation free energies based on dispersion and electrostatics. Our approach, which is purely physical and involves no fitting of parameters to data sets, executes as fast as the popular Generalized Born solvation model, but with substantially improved accuracy in prediction of experimental solvation free energies. Also, the structure of our model means that improvements in simulation forcefields will improve our results as well. All of this comes without any artificial parameter adjustments; our model's properties are the same as those used directly in molecular dynamics. Our model's energetic accuracy and detailed structural information have wide-ranging implications for molecular modeling research.