Multivariate data analysis of fatty acid content in the classification of olive oils developed through controlled crossbreeding

The fatty acid (FA) composition of 540 Tunisian virgin olive oil hybrids (VOO) were classified by principal component analysis (PCA). Pearson correlation between FA variables revealed an inverse association between C18:1 and C18:2; C18:1 and C16:0, while C16:0 and C16:1 were positively correlated. PCA yielded five significant PCs, which together account for 79.95% of the total variance; with PC1 contributing 36.84% of the total. Eigenvalue analysis revealed that PC1 was mainly attributed to C18:1, monounsaturated fatty acids (MUFA) and the ratios oleic/linoleic (O/L) and monounsaturated fatty acids/polyunsaturated fatty acids (MUFA/PUFA); PC2, by C16:0, saturated fatty acids (SFA) and the palmitic/linoleic ratio (P/L); PC3 by C18:2 and C22:0, PC4 by C18:0 and PC5, by C17:1. Then, PCA analysis indicated that in addition to C16:0, C18:0, C18:1, C17:1, and C22:0, MUFA, SFA and the ratios O/L, P/L and MUFA/PUFA were determined to be the main factors responsible for the olive oil hybrids discrimination. © AOCS 2011.