A Recalculation of Quantitative Structure Chromatographic Retention Time Relationships on Natural Phenols and Sterols Found in Olive Oil

A new approach to predict the chromatographic retention times for a group of natural phenols and sterols on olive oil is presented. A QSPR treatment from a initial set of ca. 450 molecular descriptors allows us to obtain the following empirical functions: a) a model relating the retention index of natural phenols in the olive oil with five molecular descriptors; b) two models which permit to calculate the retention index on two diferent columns (SE-54 and SE-52) for natural sterols (trimethyl-silyl ethers) from the olive oil with four quantum-chemical descriptors. In all cases, the correlations coefficients of the empirical functions are higher than 0.99, and the mean errors range between 0.01 and 0.37%. The new models found with the descriptors generated with the HYPERCHEM 4.0, AMPAC 6.7 and CODESSA 2.3 programs are the preferable for predictions because of their highest R value and lowest error percentage if compared with previously reported QSRR models.