A model for estimation of activity for efavirenz analogues with the K 103N mutant of HIV-1 RT was proposedin this work. The new model is predictive that requires only variable connectivity indices in the calculations. Compared with the existing models based on Monte Carlo simulation, the new model is easier to apply with better predictive accuracy, giving an r2 of 0.85, cross-validated Q2 of 0.82 and average error of only 0.23kcal/mol for the 47 efavirenz analogues concerned. This work also demonstrates that variable connectivity indices, a class of not widely recognizedstructural descriptors, are quite useful in the QSAR studies in the fields of pharmaceutics and biochemistry. These topological indices may play a more important role in these fields in future.

The BeHeþn clusters with n=1-12 were investigated by ab initio calculations at the level of MP2 (full)/6-311+G (2d, 2p). All the most stable geometric structures were obtained. It is found that the first solvation shell around the beryllium ion core is completed at n=12, and the mechanism of cluster forming can be explained in terms of the solvated ion core model. The effects of basis set superposition error (BSSE) correction and zero-point vibrational energy (ZPE) correction to the energies were considered in the calculations.

A SAFT equation of state (EOS) combined with eight mixing rules was used to evaluate the capability of the SAFT approach for modeling the solubility of solids in supercritical fluids (SCFs). The results show that the SAFT approach gives good correlative accuracy in general, and the results are satisfactory when the three-parameter mixing rules are used, where the average absolute deviation of solid solubility in the SCF phase is normally smaller than 10%. The present work shows that the SAFT approach can be used to model solid-SCF equilibria, which gives slightly better correlative accuracy than the cubic EOSs.

A correlation for estimation of the aqueous solubility of organic compounds that is based on a training set of 120 chemicals is proposed. The new model proposed is predictive and requires only molecular connectivity indices in the calculations. The calculated results of the new model are comparable to those from the existing general solubility equation (GSE) and the Klopman-Zhu models. The new model was also applied to a testing set of 80 compounds, and the predictions show that the new model is reliable with good predictive accuracy. Because the new model does not require any experimental physicochemical properties in the calculation, it is simple and easy to apply. This work shows again that molecular connectivity indices are useful structural descriptors in quantitative structure-property (QSPR) studies in pharmaceutical research.

An open-cell model equation of state has been proposed, which relaxes the restrictions of the cell theory and allows more flexibilities in describing the structure of liquids. To test its capability, this equation has been applied to a variety of liquids, including 17 heavy molecular weight hydrocarbons, 10 organic solvents, and 15 polymers, in particular. The calculated density grand average AADp for the equation proposed here are 0.062, 0.057, and 0.054 per cent against the results for the equation of state by Simha and Somcynsky, 0.081, 0.093, and 0.072 percent for the above three groups of liquids, respectively. Good agreement between our calculated results and experimental data indicates that our model is an improvement over the Simha and Somcynsky equation of state, which has been evaluated as one of the best equations. Moreover, this equation can be easily extended to describe the properties of polymer solutions and polymer blends, which will be addressed in our subsequent paper.