A Quantitative Structure-activity Relationship (QSAR) Study of the Anti-tuberculosis Activity of Some Quinolones

Abstract

Aims: To use QSAR methodology in developing mathematical models for predicting the in-vitro anti-tuberculosis activity of some quinolone compounds against Mycobacterium smegmatis. Study Design: A quantitative structure-activity relationship (QSAR) study on a set of thirty-four 8-methylquinolones was performed. The genetic algorithm (GA) and multiple linear regression analysis (MLRA) were used to select the descriptors and to generate the correlation models that relate the structural features to the biological activities. Place and Duration of Study: Physical Chemistry Laboratory, Ahmadu Bello University Zaria, Nigeria. Between March 2015 and July 2015. Methodology: The molecular structures of the compounds were optimized at the Density Functional Theory (DFT) level of theory using  the standard pople’s basis set 6311G* and the Becke’s three-parameter hybrid functional with LYP correlation functional  (BLYP/6311G*). Several molecular descriptors (i.e., features) were computed mainly using the Padel software for the thirty four compounds, genetic algorithm was used to choose the most relevant descriptors among the several calculated descriptors. Multiple linear regression analysis was used to develop linear model for predicting the biological activity. Results: The most robust model was found to have R2 = 0.9184. The robustness of the chosen model was further tested using the leave-one-out (LOO) cross validation procedure (Q2 LOO = 0.84987) and the external validation procedure (R2Pred =0.79343) as well as Y-randomization. Leverage approach was used to establish the applicability domain of the model. Conclusion: The predictive ability of the model was found to be satisfactory and could aid in the design of similar group of anti-tuberculosis drugs.