Application of Topological Descriptors in QSAR Modeling: Substituted Hydrazones Used As a Model System

Background: QSAR study of p-substituted aromatic hydrazones was performed to estimate the quantitative effects of selected topological descriptors on their antimicrobial activity. None of the hydrazones inhibited the growth of the Aspergillus spp., while the data obtained with regard to the antifungal activity of the compounds against Candida utilis were insufficient to develop reliable statistical QSAR models. Therefore, the investigation was focused on developing QSAR models for predicting the antibacterial activity of the compounds against Bacillus subtilis. Methods: A set of substituted hydrazones were tested for their in vitro growth inhibitory activity against Candida utilis, Bacillus subtilis and Aspergillus niger and the diameter of the inhibition zone (mm) was measured. The inhibitory activity data, determined in μg/mL, were transformed to the negative logarithms of molar MICs (log1/CMIC). Using Marvinsketch software package, 28 topological descriptors were calculated. Statistical parameters, such as R2, Sd, F-test, R2 adj, Q, SPRESS, PSE and Q2, were used to test the quality of the developed two-, three-, four-parametric and higher QSAR models. Results and Conclusion: Statistical evaluation of the data used to test the quality of the obtained QSAR models indicated that the two-parametric model involving the descriptors Atom Count (AC) and Maximal Projection Area (MAPA) was statistically significant when all the statistical parameters were summarized. The two parameters, AC and MAPA, had opposite input in modeling the antimicrobial activity of the selected hydrazones against Bacillus subtilis.