2025, Vol. 6, Issue 2, Part B
QSPR modelling based on topological indices and regression analysis for predicting the physicochemical properties of herbal glycosides
Author(s): S Ranjitham, OV Shanmuga Sundaram and G Priyadharsini
Abstract: Quantitative Structure Property Relationship (QSPR) modeling is shown to be an interesting method to predict the physicochemical properties of herbal glycosides that are compounds known in terms of their medicinal importance, though in many cases, they may not have a lot of experimental data published about them. The paper aims at investigating the use of degree-based topological indices as molecular descriptors in the construction of regressions to predict several important physicochemical properties, including boiling point, flash point, polarizability, surface tension, and molar volume, in a selection of herbal glycosides. Topological indices like First and Second Zagreb, Forgotten, Yemen and sum-connectivity were used in doing a linear regression analysis. These statistical tests indicated that boiling point, polarizability, and molar volume had the highest correlations and prediction accuracies with some of the topological indices (R2 > 0.90, p<0.001), specifically the Forgotten and the Yemen one. Conversely, the flash point and surface tension predictions exhibited poorer model performance suggesting a certain degree of lack of degree-based descriptors describing these properties. These findings indicate that topological indices can be used as effective predictors of some important properties of herbal glycosides, and in silico search and property estimation can be performed very efficiently with them. It has also been established that QSPR methodologies have the potential to contribute to the development of phytopharmaceutical development, though more extensive sets of descriptors and the application of more sophisticated model development methods are required to handle more complex property predictions. Future works must be concentrated on the enlargement of set of data and the further development of additional computational methods that will allow to increase the robustness and range of QSPR models on natural product research.
DOI: https://doi.org/10.22271/math.2025.v6.i2b.236
Pages: 227-232 | Views: 113 | Downloads: 29
Download Full Article: Click Here
How to cite this article:
S Ranjitham, OV Shanmuga Sundaram and G Priyadharsini. QSPR modelling based on topological indices and regression analysis for predicting the physicochemical properties of herbal glycosides. Journal of Mathematical Problems, Equations and Statistics. 2025; 6(2): 227-232. DOI: 10.22271/math.2025.v6.i2b.236
