Predicting antimicrobial activities of benzimidazole derivatives

Abstract

A data set of 31 benzimidazole derivatives with antibacterial activities against gram-positive (e.g. Staphylococcus aureus, methicillin-resistant S. aureus (MRSA) and Bacillus subtilis) and gram-negative (e.g. Escherichai coli) bacteria as well as antifungal (e.g. Candida albicans) activity were used for quantitative structure–activity relationship (QSAR) study. Compounds were characterized by quantum chemical descriptors calculated at B3LYP/6-31G(d) level and molecular descriptors generated using Dragon software. The QSAR model was constructed using multiple linear regression method and sampled by leave-one-out cross-validation and the predictive performance can be deduced from the correlation coefficient (r), which was in the range of 0.4194–0.8485 while root mean square error (RMSE) was in the range of 0.039–0.3247. Particularly, QSAR model against S. aureus, MRSA (clinical isolates) and E. coli provided good predictive ability with r values of 0.8485, 0.8443 and 0.833, respectively, and RMSE values of 0.2662, 0.2488 and 0.039, respectively. The QSAR models demonstrated high potential for the rational design of novel benzimidazole derivatives with potent antimicrobial activities.