Molecular simulations can support the rational drug design, as they can be used to determine solvation free energies and thus relative solubilities. However, common classical molecular models (force fields) were not primarily generated to describe solvation free energies and important polarization effects are not considered. Thus, based on the General Amber Force Field (GAFF), a new model, GAFF/IPolQ-Mod+LJ-Fit was developed, which implicitly accounts for polarization effects using the IPolQ-Mod approach. New model parameters to describe Lennard-Jones interactions were fitted to 357 solvation free energies and 384 densities of pure substances. For validation, 237 relative solubilities of complex drug molecules and 100 densities of binary mixtures were considered. Compared to the standard GAFF model as a reference, the RMSD deviations by the new model are reduced by 63% and 36%, respectively, so that it allows for significantly improved predictions of thermophysical properties.
Original article
Mecklenfeld, A. and Raabe, G. GAFF/IPolQ-Mod+LJ-Fit: Optimized force field parameters for solvation free energy predictions. ADMET & DMPK 8 (3) (2020) 274-296; doi:http://dx.doi.org/10.5599/admet.837.