An Investigation on the QSAR Modeling of Carfilzomib Derivatives Using Monte Carlo Method and Novel Modelling-optimization Approach

Document Type : Full Paper


Department of Chemistry and Chemical Engineering, Rasht Branch, Islamic Azad University, Rasht, Iran



The activity of the 25 different Carfilzomib derivatives was estimated using multiple linear regression (MLR), artificial neural network (ANN), and genetic algorithm(GA) and simulated annealing algorithm (SA) and Imperialist Competitive Algorithm (ICA) as optimization methods.
The obtained results from MLR-MLR, MLR-GA, SA-ANN and GA-ANN techniques were compared and for combinations of modelling-optimization methods observed root mean sum square errors (RMSE) of 0.290, 0.0482, 0.0294, 0.0098 in gas phase, respectively (N=25).
A high predictive ability was observed for the MLR-ICA model with the best number of empires/ imperialists (nEmp=50 ) and nEmp=100 with root-mean-sum-squared error (RMSE) of 0.00996 in gas phase.
From the MLR-ICA method, it was revealed that RDF 075m, MATS1m, F04[N-O], O-059, F09[C-O] and Mor21p are the most important descriptors. From Monte Carlo simulations, it was found that the presence of double, absence of halogens, oxygen connected to double bond, sp2 carbon connected to double bond, double bond with ring, branching, nitrogen are the most important molecular features affecting the biological activity of the drug.
It was concluded that simultaneous utilization of MLR-ICA, GA-ANN and Monte Carlo method can lead to a more comprehensive understanding of the relation between physico-chemical, structural or theoretical molecular descriptors of drugs to their biological activities and facilitate designing of new drugs.