Predicting Degradation Half-life of Organophosphorus Pesticides in Soil Using Three-Dimensional Molecular Interaction Fields

Abstract

A simple and strong model, based on an alignment independent three-dimensional quantitative- structure activity relationships (3D-QSAR), is developed for prediction of degradation half-life (DT50) of 47 organophosphorus pesticides in soil. Molecular descriptors derived from 3D molecular interaction fields (MIF) were calculated using the GRIND methodology. Fractional factorial design (FFD) applied to feature selection and modeling of the relationship between selected descriptors and DT50 data was achieved using partial least squares regression. Validation and reliability of the obtained model were checked by the prediction of external test set cross-validation and chance correlation. The value of the determination coefficient (R2) was 0.817 for leave-one-out cross-validation procedure. The R2 values for the training and test sets were 0.951 and 0.893 respectively. The obtained model confirmed that size and shape of the molecules as well as hydrophobic interactions are the main parameters influencing the degradation half-life of organophosphorus pesticides in soil.