QSAR study of antiproliferative drug against A549 by GA-MLR and SW-MLR methods

Abstract

Quantitative structure-activity relationship (QSAR) is the most extensively used computational methodology for analogue-based design. In this research, QSAR model was used to predict antiproliferative properties of 4-(2-fluorophenoxy) quinoline derivatives against A549(human lung adenocarcinoma). For this purpose, we used the multiple linear regressions (MLR) for the construction of a model to predict drug activity and Stepwise (SW) and genetic algorithm (GA) methods used to build the model. The data were selected from 31 molecules with specific pharmacological activity. They were divided into two sets train and test data. The resulting model was tested using statistical methods such as external test set and cross-validation to ensure its authenticity. The results showed that GA-MLR approach had good predictive power and higher data rates compared with SW-MLR (Q2LOO = 0.877, R2Train =0.933). The results obtained in this study can be used to design drugs with higher performance and pharmacological activity to treat lung cancer.

Keywords

• lung cancer
• quinoline derivative
• multiple linear regressions
• genetic algorithm