Applying quantitative structure-toxicity relationship (QSTR) model to predict the toxicity of pesticide carbamates using computation methods and molecular descriptors

Subject Areas : entomology and othea arthropods

Seyedeh Azadeh  Moosavi ¹ (, Department of Chemistry, Arak Branch, Islamic Azad University, Arak, Iran)
Esmat  Mohammadinasab ² (Islamic Azad University, Arak Branch, Iran)
Tahereh  Momeni Isfahani ³ (, Department of Chemistry, Arak Branch, Islamic Azad University,, Arak, Iran)

Received: 2022-11-30 Accepted : 2023-03-03 Published : 2023-02-20

Keywords:

Abstract :

In this study, we performed quantum mechanics computation at density function theory level with 6-31G* basis set to construct a quantitative structure-toxicity relationship (QSTR) model for predicting lethal dose (LD50) pesticide carbamates derivatives. The best molecular descriptors were selected using genetic algorithm (GA) by MATLAB software. Then, we studied the relationship between the selected descriptors and the logLD50 of carbamate derivatives using backward-stepwise multiple linear regression (BW-MLR) and backpropagation artificial neural network (BP-ANN) models. The RDF010e, WW, and R3e descriptors were applied for modeling the GA-BWMLR and GA-BPANN models. The comparison of results illustrated that the R2 and Q2 of GA-BPANN model for all set were significantly higher than the GA-BWMLR model. The GA-BPANN model was more accurate with lower mean square error (MSE), root-mean-square error (RMSE), standard error of prediction (SEP), and absolute average deviation (ADD) values of data set for predicting the LD50 of studied carbamates.