QSPR Analysis of Kidney Infection (Pyelonephritis) Drugs by Entropy Graphs Weighted with Topological Indices, and MATLAB Programming

The entropy of the chemical graph primarily measures the complexity of chemical structures. In this study, Shannon's entropy approach was utilized to calculate entropies for chemical graphs with degree-based topological indices, predicting the physicochemical correlation capability of 12 drugs for kidney infection and its implications for physicochemical aspects. Entropy can help identify important correlations between the structure and function of compounds, which in turn can aid in developing QSPR models. A computer-based computing technique and algorithm were employed to simplify calculations and data analysis. Degree-based topological indices, along with the entropy graphs weighted with indices, were calculated using the MATLAB program and used to measure entropy. This research aims to predict the physicochemical characteristics of these drugs based on the entropy of topological indices. Several regression models were used to analyze drugs' quantitative structure-property relationship (QSPR) and predict their physicochemical properties. Effective predictors were identified in this study, and optimal equations were introduced to evaluate the physicochemical characteristics based on the highest correlation coefficient and Fisher's statistical test. Interestingly, in some cases, the entropy of degree-based indices showed a stronger correlation with drug characteristics. In conclusion, the optimal equations for estimating molar refractivity (MR) and polarizability (P) using the ENTRR-1/2, ENTSC, and ENTABC indices incorporate compound, growth, and exponential factors. Meanwhile, the optimal equations for estimating molar volume (MV) with the ENTF index are quadratic and cubic.