Fatty Acid Profile of Table Olives and Its Multivariate Characterization Using Unsupervised (PCA) and Supervised (DA) Chemometrics ANTONIO LO Ä PEZ,ALFREDO MONTAN ˜ O,* PEDRO GARCI ÄA, AND ANTONIO GARRIDO Food Biotechnology Department, Instituto de la Grasa (CSIC), Avenida Padre Garcı ´a Tejero 4, Seville, Spain The fatty acid composition of 67 commercial presentations of table olives was determined. The most abundant fatty acids, in decreasing order of presence, were C18:1, C16:0, C18:2 n-6, and C18:0. The ranges, expressed as grams of fatty acids per 100 g of edible portion, for the different nutritional fractions were as follows: saturated fatty acids, 2.07-5.99; monounsaturated fatty acids, 5.67- 19.42; polyunsaturated fatty acids, 0.52-3.87; and trans-fatty acids, 0.08-0.44. Principal component analysis of the matrix of the fatty acid composition led to the deduction of new factors. The first accounted for 55.10% of the total variance and was mainly related to C16:10, C18:0, C20:0, C22:0, C24:0, C18:1, C18:1t, and C20:1. The second factor accounted for 10.33% of variance and was related to C16:1 and C18:2 n-6. They did not permit differentiation among elaboration types or cultivars. However, discriminant analysis was successfully applied for this objective. The fatty acids that most contributed to discriminate among elaboration styles were C17:1, C18:1, C16:0, C17:0, and C18:0 (function 1) and C17:0, C17:1, C20:0, C16:0, C18:1, and C24:0 (function 2). In the case of cultivars, they were C20:0, C18:1, C17:1, C18:2 n-6, C18:1t, and C18:2t (function 1); C18:2 n-6, C18:1, C16: 0, C20:0, C18:0, and C18:2t (function 2); and C17:0, C18:3 n-3, and C17:1 (function 3). Results from this study have shown differences among the fatty acid composition and fat content of the diverse commercial presentations of table olives, which can be applied in predictive and classification discriminant analysis. KEYWORDS: Fatty acids; table olives; saturated fat; monounsaturated fat; polyunsaturated fat; trans fat; principal component analysis; discriminant analysis INTRODUCTION The fruits of the olive tree are mainly used for the extraction of olive oil, although about 20% of them are prepared as table olives. Table olive production reached a total of 1 600 000 tonnes in the 2003/2004 season, Spain being the main producer and exporter with about 500 000 and 250 000 tonnes, respec- tively, during this period (1). The proximate composition of this product with respect to the main compounds (total fat, total carbohydrates, fiber, etc.) is known and has shown that, after moisture, fat is the most outstanding component (2). However, studies related to the composition of the fat from table olives are scarce (3, 4), and the comparative effect of the different processing styles or cultivars on its content has never been studied. On the contrary, there are numerous references in the literature related to the composition of olive oil. Chemometric techniques have been used extensively in olive oil with diverse objectives (5-8). Unsupervised methods reveal relationships between oil samples and, at the same time, among analytical data, to produce a clustering of variables and samples into distinct groups. In supervised methods, a set of data describing oils of known origins is used to construct models that are then applied to classify unknown oil samples into a priori established groups, either geographically or by cultivar. Lee et al. (9) used the fatty acid (FA) composition in vegetable oils together with principal component analysis (PCA) and discriminant analysis (DA) to differentiate among sesame, perilla, soybean, corn germ, canola, rapeseed, olive, and coconut oils and proposed these techniques to detect adulterations. FA analysis and chemometrics have also been used in other food or vegetable products. Abrodo et al. (10) studied the FA composition to differentiate between traditional and controlled cider fermentation. The use of PCA, DA, soft independent modeling of class analogy (SIMCA), and partial least squares (PLS) allowed the authors to typify fermented apple products on the basis of fermentation technology, lauric and palmitic acids being the most relevant variables for classification purposes. FAs from seeds of Pinus pinea L. were studied by PCA and showed that the low genetic diversity revealed by acid composi- * Author to whom correspondence should be addressed (telephone +34 95 4691054; fax +34 95 4691262; e-mail amontano@cica.es). J. Agric. Food Chem. 2006, 54, 6747-6753 6747 10.1021/jf0612474 CCC: $33.50 © 2006 American Chemical Society Published on Web 08/08/2006