QSPR modeling of physicochemical properties of SSRI and SNRI antidepressants using advanced molecular graph descriptors
TL;DR
The results of the present study indicate that the neighborhood degree-based topological indices are efficient tools for the development of QSPR models and can be applied in the rational design, screening and optimization of novel antidepressant drugs.
Abstract
The Quantitative Structure–Property Relationship (QSPR) modeling is a successful computational technique that is capable of predicting physicochemical properties on the basis of the molecular structure. Twenty-five commonly used antidepressant drugs were used in the present study and a complete QSPR analysis was carried out by using neighborhood degree-based topological indices obtained from chemical graph theory. Molecular graphs, with the hydrogen atoms suppressed, were calculated nine topological descriptors; the correlations between these descriptors and eight measured physicochemical properties (boiling point, molar volume, molar refractivity, molecular weight, heavy atom count, molecular complexity, enthalpy of vaporization, and melting point) were computed. There were development and statistical assessment of three regression models (linear, quadratic and logarithmic) using coefficient of determination (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$R^{2}$$\end{document}), F statistic, p-value, and standardized residuals. The comparative analysis showed that for the boiling point, the quadratic regression model is generally the most accurate predicting model, with a maximum \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$R^{2}$$\end{document} value: 0.887 for the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ND_1(G)$$\end{document} descriptor. Moreover, the two most informative neighborhood descriptors that consistently emerged as the most informative neighborhood descriptors for predicting several important physicochemical properties were \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ND_1(G)$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ND_5(G)$$\end{document}. The developed regression equations are reliable and efficient predictive models, and do not require expensive experimental measurements during the early stages of drug discovery and require only molecular topology. The results of the present study indicate that the neighborhood degree-based topological indices are efficient tools for the development of QSPR models and can be applied in the rational design, screening and optimization of novel antidepressant drugs.