Prediction of octanol – water partition coefficient of quinoline alkaloid insecticides using Molecular Descriptors and MLR Method

Quinoline alkaloids and their derivatives have wide medical and agricultural applications. In this research, a quantitative structure- property relationship (QSPR) has been employed to predict the octanol-water partition coefficient (logP) of 76 quinoline alkaloid camptothecin (CPT) derivatives as antitumor potencies using GA-MLR method and molecular descriptors. The Gauss View 05 software was used for drawing chemical structure of the studied compounds. The geometry optimizations of the studied compounds were done by the Gaussian 09W software at B3YLP density functional theory (DFT) with 6-311G (d,p) basis set. Molecular descriptors for each of optimized structures were calculated by Dragon software in different category. In order to reduce and select the best descriptors, the Genetic Algorithm technique and stepwise multiple linear regression method was used. The pearson coefficient correlation (PCC) and the variance inflation factor (VIF) statistics were used to test the multicollinearity of descriptors in the best model. The different types of internal and external validations were used to evaluate predictive model performance. The best QSPR model is obtained with R2 value of 0.901, Q2LOO =0.919, and RMSE=0.706.     The results of statistical parameters and validations of the GA-MLR model generated were found to be satisfactory. The model revealed that octanol-water partition coefficient of CPT derivatives is influence by ATS8e (2D-autocorrelation) descriptor. This information could be used to design novel quinoline alkaloid camptothecin (CPT) derivatives as insecticide agents.