Comparison of aqueous solubility estimation from AQUAFAC and the GSE

The GSE (General Solubility Equation) and AQUAFAC (Aqueous Functional Group Activity Coefficients) are two empirical models for aqueous solubility prediction. This study compares the aqueous solubility estimation of a set of 1642 pharmaceutically and environmentally related compounds, using the two methods. The average absolute errors in the solubility prediction are 0.543 log units for AQUAFAC and 0.576 log units for the GSE. About 88.0% of the AQUAFAC solubilities and 83.0% of the GSE molar aqueous solubilities are predicted within one log unit of the observed values. The marginally greater accuracy of AQUAFAC is due to the fact that it utilizes fi tted- parameters for many structural fragments and is based on experimental solubility data. The GSE on the other hand is a simpler, non-regression based equation which uses two parameters for solubility prediction. Published by Elsevier B.V.