SAR AND QSAR IN ENVIRONMENTAL RESEARCH, cilt.16, sa.5, ss.443-459, 2005 (SCI İndekslerine Giren Dergi)
QSPR models for water solubility (S), n-octanol/ water partition coefficient (K-OW), and Henry's law constant (H) for polychlorinated dibenzo-p-dioxins (PCDDs) and dibenzo-p-furans (PCDFs) and phthalates have been established based on two different sets of parameters. Those parameters were topology based characteristic root index (CRI) and three semi-empirical molecular descriptors, namely - energies of the highest occupied and the lowest unoccupied molecular orbital (E-HOMO and E-LUMO), and dipole moment (mu). The best fit equation found by 'forward multiple linear regression' showed that the topology based CRI was the most important parameter for the modelling of solubility and n-octanol/ water partition coefficient. For n-octanol/ water partition coefficient a two-parameter equation including the CRI and E-HOMO with a correlation coefficient of r = 0.992 was obtained whereas a three-parameter equation for solubility and Henry's law constant including the CRI, E-LUMO and mu with a correlation coefficient of r = 0.986 and r = 0.933 was obtained, respectively. E-HOMO and mu didn't appear in the same model because of the collinearity. The results of modified jackknife tests indicated that the three models were statistically robust. Mean deviation of calculated values from experimental data amounted to 0.27, 0.17, and 0.28 log units for the three properties mentioned. The developed models have been used to predict the S, KOW and H of compounds not included in the training sets.