QSAR study and molecular modeling of pleuromutilin derivatives to predict antibacterial activity: Chemometrics and...

Quantitative structure-activity relationship (QSAR) as one of the most important fields of chemometrics provides useful information in pharmaceutical chemistry, drug design, toxicology, and other related areas. In this study, a QSAR model was developed employing image pixels of molecular structure as descriptors to describe the antibacterial activity of pleuromutilin derivatives. Accordingly, a dataset of 54 Pleuromutilin derivatives was collected from the literature, plotted, and transformed into pixels using MATLAB software to generate the desired model. The data were pre-processed through mean-centering and scaling to optimize pixel selection. The suitable descriptors were then extracted using principal component analysis (PCA) and the final model was built followed by partial least squares (PLS) multivariate calibration. The developed QSAR model was validated and novel compounds with antibacterial properties were suggested. In addition, molecular docking studies were utilized to investigate inhibitory activities of proposed new compounds against ribosomal protein (PDB ID: 3G4S). The findings showed high potential of proposed compounds to form strong interactions with key active site residues, confirming their good antibacterial activity and the validity of constructed QSAR model for drug prediction.