QSRR study of Gas Chromatographic Behavior of Selected Pesticides in Green Tea Samples

Different classes of pesticides are used during production of various tea products. Beside their health impact on the consumers, these pesticides are likely to be released into the different natural media and affect the exposed environment. To help predict chromatographic behavior of these pesticides, the gas chromatographic retention behavior of 57 of these pesticides were modeled using quantitative structure-retention relationship approach. The most important descriptors encoding structural and topological properties of the studied compounds were showed to be in a linear relation with their retention time. A stepwise variable selection strategy in MLR modeling resulted in models with acceptable R2 values of which a model based on five molecular descriptors was selected to compromise between high R2 values and low variable numbers. The selected descriptors were VRD2, X1sol, Rww, MLOGP and More09e. The model was tested for its prediction capability, by examining a prediction set of randomly selected compounds (10 pesticides) and the average of the prediction error was used as the criteria. The model, regarding its simplicity, was successful in predicting the retention times of the proposed set of compounds.