Estimation of aqueous solubility for a diverse set of organic compounds based on molecular topology

An accurate and generally applicable method for estimating aqueous solubili ties for a diverse set of 1297 organic compounds based on multilinear regre ssion and artificial neural network modeling was developed. Molecular conne ctivity, shape, and atom-type electrotopological state (E-state) indices we re used as structural parameters. The data set was divided into a training set of 884 compounds and a randomly chosen test set of 413 compounds. The s tructural parameters in a 30-12-1 artificial neural network included 24 ato m-type E-state indices and six other topological indices, and for the test set, a predictive r(2) = 0.92 and s = 0.60 were achieved. With the same par ameters the statistics in the multilinear regression were r(2) = 0.88 and s = 0.71, respectively.