QSPR study on the estimation of solubility of drug-like organic compounds: a case of barbiturates.

The paper describes a method for the estimation of solubility (log S) of a series of 45 barbiturates employing 26 molecular descriptors. The molecular descriptors used being distance-based topological indices, information indices, valence connectivity index, shape indices, first-order Randic index. In addition, an indicator parameter was also used. The regression analysis has shown that an R(2) value of 0.885 was obtained in multi-parametric regression analysis. The results are discussed critically using a variety of statistical parameters. The predictive powers of the models were discussed by using the method of cross-validation. We observed that results obtained using SPSS and NCSS software are identical.