Optimized Parameters for the Drude Polarizable Force Field for Small Organic Molecules

Abstract

Meaningful efforts in computer-aided drug dsign (CADD) need accurate molecular mechanical force fields to quantitatively characterize protein-ligand interactions. The most widely used strategy to ascribe force field parameters for new compounds apply pre-defined tabulated values following knowledge-based rules. Two widely used approaches following this strategy are the General Amber Force Field (GAFF) and the CHARMM General Force Field (CGenFF). While these approaches are extremely useful, an important limitation of using pre-tabulated parameter values is that they may be inadequate in the context of a specific molecule. To resolve this issue, we previously introduced the General Automated Atomic Model Parameterization (GAAMP) to automatically generate the parameters of atomic models of small molecules using the results from ab initio quantum mechanical (QM) calculations as target data [DOI:10.1021/ct4003477]. The GAAMP protocol uses QM data to optimize the bond, valence angle, and dihedral angle internal parameters, and atomic partial charges. However, the Lennard-Jones (LJ) 6-12 parameters cannot easily be refined from QM data and are unchanged from the initial atom types assignments severely limiting the achievable accuracy by these models. Here, we systematically optimize the LJ parameters of the Drude force field to reproduce experimental neat liquid densities and enthalpies of vaporization for a large set of compounds, covering a wide range of chemical functionalities. The optimization is carried out by constructing an objective function and exploiting the analytical gradient information within an optimization framework. The new set of LJ parameters will help to improve the accuracy of the computed physical properties of diverse small organic molecules. These developments will provide the necessary tools to widely apply drude polarizable force field model for drug discovery endeavors.