engleski

QSAR modeling for toxicity of pesticides to aquatic organisms using easily interpretable descriptors

A small fraction of the pesticides consumed in agricultural or urban settings finish up moving with surface runoff into streams, rivers, and lakes, leaching to the groundwater systems or volatilizing to the atmosphere. Hence, toxicity estimation of pesticides is important. In the present work, QSAR (Quantitative Structure- Activity Relationship) analysis for toxicity of a dataset of pesticides comprising organophosphates, triazine, has been performed [1]. A good number of molecular descriptors were calculated using PaDEL and a new in-house built PyMOL plugin (PyDescriptor) followed by extensive objective and subjective feature selection to avoid redundant descriptors. For model building, the dataset was divided into training (80%) and test (20%) sets. A QSAR model built using three easily interpretable descriptors was subjected to extensive internal and external validation. The QSAR model is statistically robust with R2 = 0.872, Q2 = 0.844, CCCex = 0.845. The analysis revealed that lipophilicity, frequency of occurrence of hydrogen within 3 Å from phosphorus, and the presence of two benzene rings with –CH2– group as linker have good correlation with the toxicity of the pesticides.

QSAR ; pesticides ; toxicity

nije evidentirano