Abstract :
In this paper, the quantitative structure activity-relationship (QSAR) of the CCR2b receptor inhibitors was scrutinized. Firstly, the molecular descriptors were calculated using the Dragon package. Then, the stepwise multiple linear regressions (SW-MLR) and the genetic algorithm multiple linear regressions (GA-MLR) variable selection methods were subsequently employed to select and implement the prominent descriptors having the most significant contributions to the activities of the molecules. A combined data set including numerical values of inhibition activity data (IC50) of 103 CCR2b receptor derivatives was adopted for our simulations. This study revealed that both SW-MLR and GA-MLR methods consisted of six molecular descriptors. The adopted descriptors belong to topological, charge, RDF and atom-centered fragments classes. A comparison of results by the two methodologies indicated the superiority of GA-MLR over the SW-MLR method. The authenticity of the proposed model (GA-MLR) was further confirmed using the cross-validation, validation through an external test set and Y-randomization.
