Predicting Binding Affinity of Some Imatinib Derivatives as BCR-ABL Tyrosine Kinase Inhibitors Based on Monte Carlo Optimization

Introduction: The Imatinib drug is used to treat blood cancer by inhibiting the BCR-ABL tyrosine kinase enzyme, which prevents the proliferation of cancer cells. Materials & Methods: In order to predict the binding affinity of 555 compounds of imatinib derivatives as ABL-BCR tyrosine kinase inhibitors, quantitative structure-activity relationship (QSAR) modeling was performed using the Monte Carlo method. The data were randomly divided into four series, including training, invisible training, calibration, and validation sets, as well as they were randomly repeated three times. Results: The results of three random divisions indicated reliable models for predicting the set of external tests with correlation coefficient (R2) and cross-validation correlation coefficient (Q2) in the range of 0.8575-0.8775 and 0.7620-0.7793. Consequently, the obtained models help identify hybrid descriptors for increasing or decreasing binding affinity (Ki) as BCR-ABL tyrosine kinase inhibitors. The mechanical interpretation of the model is given in the form of a report of descriptors that decrease and increase pKi, as well as examples of these descriptors. Conclusion: The results reveal that the designed models can be considerably effective in estimating the biological effect of imatinib derivatives proposed by researchers and medicinal chemists. Therefore, it is possible to predict its possible biological effects by spending less time and money before conducting in vitro or in vivo experiments.