Abstract
A set of 24 descriptors consisting of quantum and chemical descriptors have been used to model binding constant (logK) of the benzene sulfonamides to human CAII. Simple as well as multiple regression have indicated that MNC (most negative charge) is the most dominating parameter to be used in modeling log K. Excellent results are obtained in multi-parametric regression. The results are critically discussed using a variety of statistics, which indicated that the hydrophobic term (log P) is not essential to yield excellent models.
Keywords: QSAR, Quantum parameter, Hydrophobicity, Regression analysis, Chemical parameter