Development of quantitative structure-property relationship models for early ADME evaluation in drug discovery. 1. Aqueous solubility

A simple QSPR model, based on seven 1D and 2D descriptors and artificial ne ural network, was developed for fast evaluation of aqueous solubility. The model was able to predict the molar solubility of a diverse set of 1312 org anic compounds with an overall correlation coefficient of 0.92 and a standa rd deviation of 0.72 log unit between the calculated and experimental data. Considering the fact that the estimated uncertainty of the experimental da ta is no less than 0.5 log unit, the results demonstrate that carefully cho sen physically meaningful ID and 2D descriptors encode sufficient molecular information for fast and reasonably reliable prediction of aqueous solubil ity with a simple neural network. As a comparison, we calculated the solubi lity of a test set of 258 compounds, ranging from simple hydrocarbons to mo re complex multifunctional organic molecules, with a commercial program (QM PR+ version 2.0.1 of SimulationPlus Inc.) and compared the results with pre dictions from our model. Statistical parameters indicate that for small and simple organic compounds, QMPR+ outperforms our model. However for more co mplex multifunctional molecules, our model is superior.