Prediction of SAMPL-1 Hydration Free Energies Using a Continuum Electrostatics-Dispersion Model

Abstract

The SAMPL-1 hydration free energy blind prediction challenge data set includes 63 compounds that are more chemically diverse, polyfunctional, drug-like, and with examples of transfer free energies and molecular weights larger than ever before seen in previously tabulated data sets of neutral compounds. For the prospective SAMPL-1 study, we employed a continuum model including a boundary element solution of the Poisson equation to describe electrostatic solvation, a molecular surface area-based cost of cavity formation in water, and a continuum Lennard-Jones potential to account for dispersion-repulsion solute-solvent effects. For the latter contribution, continuum van der Waals atom-type coefficients were calibrated and validated on previously available hydration data sets. In the prospective study, this continuum hydration model yielded SAMPL-1 predictions highly correlated with experimental data, albeit with a slope of slightly above 0.5, suggesting a valid model but with a systematic error. Analysis of the major outliers, all overestimating the experimental hydration data, highlights a common structural theme as a possible cause of the prediction errors: densely polar and hydrogen-bond-capable structures, featuring primarily substituted (sulfon)amide groups, often in conjugated systems. By examining analog pairs within the SAMPL-1 data set, it was also noted that certain solvation trends are captured neither by chemical sense nor by our hydration model, which seem too additive. A retrospective analysis of model transferability between hydration data sets as a function of its parameters and complexity indicates that the electrostatic component of the model is fairly transferrable across data sets, but the nonelectrostatic terms are less so. For the chemical space covered in SAMPL-1, absolute prediction errors indicate that the simpler transferrable electrostatics-only model outperforms the more complex model including cavity and continuum dispersion terms. Possible directions to further improve this continuum hydration model are proposed.

Rapid Prediction of Solvation Free Energy. 1. An Extensive Test of Linear Interaction Energy (LIE)

“…It is possible that the apparent steeper surface area dependence of the cavity term for these compounds is only compensating for other factors, e.g., limitations of the molecular mechanics force-fields to accurately describe the interaction energy terms for this functional group, as previously suggested. 17 Indeed, most of the current state-of-theart solvation methods when applied to the SAMPL1 data set failed miserably on these sulfoneurea analogs, 17,[31][32][33] although some of the QM-based methods seem to provide a better agreement to experiment for some of the sulfoneurea analogs. 30,34 Total Nonpolar Component.…”

“…11 The SAMPL1 blind challenge operated by first releasing the molecules to the public followed by the release of experimental transfer free energies after a few months. A number of continuum and explicit-solvent methods were tested, most of them in the prospective mode, 17,[30][31][32] while results with two other methods were included retrospectively. 33,34 We have analyzed the sensitivity of the LIE model to several parameters that we believe are important for the accuracy of the model and for calibration of a continuum model based on these results.…”

Rapid Prediction of Solvation Free Energy. 2. The First-Shell Hydration (FiSH) Continuum Model

“…The FiSH model draws its roots from a continuum electrostatics-dispersion (CED) solvation model, 21 which we …”

“…Unlike in FiSH, where the parameters were trained on explicit water simulation, the parameters were calibrated against the experimental hydration free energy data. 21 The CED model uses a solute dielectric of 1, Born radii that were 0.9 of the AMBER van der Waals radii, and a continuum van der Waals model with 25 atom types, all of which were taken from the previous study. Since the CED parametrization lacked continuum van der Waals parameters for the iodine atom, CED predictions were not obtained for all molecules from the current data sets containing iodine.…”

“…In the RF model, the scanning of the scalar for the AMBER van der Waals radii, used as the Born radii, and solute dielectric parameters in a previous study suggest optimal values of 1.1 and 1, respectively, for acceptable and transferable prediction of smallmolecule hydration. 21 These values are then used for reaction field calculations by solving the Poisson equation using a boundary element method. Both models were used as implemented within the BRI-BEM program.…”