Assessment of Tunisian virgin olive oils via synchronized analysis of sterols, phenolic acids, and fatty acids in combination with multivariate chemometrics

Olive oil composition and connection between effective physicochemical factors characterizing accessions from different Tunisian farming sites viz. Chemlali Sfax, Chemalali Medenine, and Zalmati Medenine, located in the centre and the south of Tunisia, was probed in this study. The relationship between olive oil composition and physicochemical characteristics from different Tunisian cultivars, namely, Chemlali Sfax, Chemlali Medenine, and Zalmati Medenine, located in the centre and the south of Tunisia, was investigated using multivariate statistical analysis. Multiple linear regressions (MLR) and artificial neural network (ANN) methodologies were employed to expose hidden relationships between oxidative stability and olive oil components such as fatty acids, phenolic acids, and sterol contents. Obtained results showed not only that the selected components are dependent on geographical location and varietal origin of olive oils, but also that fatty acids (C16:1, C17:1, C18:0, C18:1, and C18:2), specific phenols (p-hydroxyphenylacetic, o-coumaric, and gallic acids), and sterols (campestanol, stigmasterol, and sitostanol) are directly implied in the oxidative stability variation. However, ANN analysis allowed to obtain more accurate models with higher robustness (R-2> 98%). The combined analytical approaches, MLR and ANNs, could be considered as an adequate experimental model to restrain the influence of olive oil components in characterization of olive oil quality. Here, we have addressed the aim of using different analytical instruments in this field and the application of chemometrics for sterols, phenolic acids, and fatty acid analysis.