2024/05/30 by Mahtal, Asmae, Toufik, Hamid, Akabli, Taoufik +3
paper · doi:10.48419/imist.prsm/rhazes-v19.49142
Bacterial infections are caused by different microorganisms and are the cause of most fatal diseases and the most widespread epidemics. The use of molecular modeling methods such as two-dimensional Quantitative Structure-Activity Relationship (2D-QSAR) studies is one of the most interesting ways to predict the antibacterial activity of molecules before their synthesis. 2D-QSAR analysis was applied to study a series of 26 molecules based on hydroxamic acid to establish a model linking antibacterial activity and the most significant descriptors using statistical methods. The dataset of the 26 compounds with their antibacterial activities (IC50 in µM) was collected; converted to the pIC50; scale and taken as dependent variables for the 2D-QSAR study. The structures of the molecules were traced using the MarvinSketch software and then optimized with the MM2 Force Field implemented in ChemBiooffice Software. The 26 molecules were randomly divided into two sets: 21 molecules for the training set and 5 molecules for testing its predictive power. the training set and 5 molecules for testing its predictive power. The descriptors were calculated using the software (MOE) and then the initial range of descriptors was reduced by performing the PLS analysis. After this analysis, four descriptors were retained. The results obtained show that the 2D-QSAR model, obtained by the PLS method has four descriptors dataset preparation that have a significant influence on antibacterial activity with R²=0.74, R²adjusted=0.70, RMCE=0.42, Fischer’s test = 14.93, Q2= 0.77, R2test= 0.87 and α = 5%. Overall, the selected 2D-QSAR model showed excellent results and would predict the antibacterial activity of new compounds.