Efficient Computation of the Total Solvation Energy of Small Molecules via the R6 Generalized Born Model

Efficient and accurate methodologies to compute solvation free energies of small molecules are relevant for many biological and industrial research areas including rational drug design. In this work we test the performance of a recently developed generalized Born method, GB_NSR6 (Aguilar et al. J. Chem. Theory Comput. 2010, 6, 3613-3639) on a common benchmark set of 504 small molecules. The computed solvation energies are compared with those obtained previously by explicit solvent models and experiment. The dominant polar component of the solvation energy is computed by GB_NSR6 with no adjustable parameters, producing a root mean square deviation (RMSD) of 0.89 kcal/mol with respect to explicit solvent (TIP3P). The relatively small nonpolar contribution is estimated using the Gallicchio et al. (J. Comput. Chem. 2005, 25, 479-499) approach. Our results show that GB_NSR6 offers a reasonable balance between efficiency and accuracy: the RMSD from the experiment of computed solvation energies is 1.2 kcal/mol, which is essentially the same as the accuracy of the much more computationally expensive explicit solvent treatment. The average computational time needed to compute the total solvation energy per molecule via GB_NSR6 is only tens of milliseconds on a commodity PC for a typical molecule of about 20 atoms. All of the software developed in this work is freely available from http://people.cs.vt.edu/onufriev/software.php.