Use of Topological Indices of Organic Sulfur Compounds in Quantitative Structure‐Retention Relationship Study

Abstract

AbstractStructure‐gas chromatographic retention index models were developed for some organic sulfur compounds at four different temperatures (60, 80, 100 and 120 °C) using only topological descriptors. At first, regression models were generated for each temperature separately with high values of multiple correlation coefficient and Fisher‐ratio statistics. The results of cross validation test using leave‐one‐out (Q2≥0.956) and leave‐two‐out (Q2≥0.953) methods showed good predictive ability of the models developed. Then, a single combined quantitative structure‐retention relationship model, added temperature as a parameter, was also developed for all the temperatures, showing good statistical parameters (R=0.991 and F=728.474). The stability and validity of the combined model were verified by both internal (Q2>0.970) and external validation (Q$\rm{ {_{prediction\;set}^{2}}}$=0.993) techniques. The results of the study indicated the efficiency of the classical topological descriptors in simultaneous prediction of retention index values of sulfur compounds at different temperatures. The topological descriptors well covered the molecular properties known to be relevant for gas chromatographic retention data, such as molecular size and degree of branching.