DUGi

Exploratory Data Analysis for Fatty Acid composition in Pig Meat

http://dugi.udg.edu/item/http:@@@@hdl.handle.net@@10256@@13656

Fat content and composition are determinant factors affecting pork production and meat quality (Wood et al., 2003). Fat composition is commonly presented as the percentage of each individual fatty acid relative to total fatty acids and then some pork quality traits are described in terms of some fatty acid percentages (see for example the review by Wood et al., 2008). Despite being compositional in nature, to our knowledge there is no reference in the literature where specific compositional data analysis methods had been applied to analyze fatty acid composition. A first objective of this contribution is to analyze fatty acid composition as compositional data. In meat quality research, it is common to analyze the effect of some influential factors (such as diet, genotype, gender, live weight or age, among others) on fat content and composition, usually the subcutaneous (SF) or the intramuscular (IMF) fats. In these studies, the aim is mostly to estimate and then test the differences among treatments for fatty acid percentages. The pattern of fatty acid deposition may differ among tissue (for instance, SF or IMF) or muscle (Kloareg et al., 2007; Duran-Montgé et al., 2008), and even between localization within a specific tissue. A second objective of this manuscript is to better know the differences between fatty acid deposition pattern between tissues, localization within a tissue, and muscles. Thus, the purpose of the present study is, first, to describe a data set of fatty acid composition collected specifically for doing research on IMF content and composition and assess the main differences among IMF of three muscles and SF using compositional data methods, and, second, to apply, as a case study, specific compositional data methods to discriminate between samples of IMF and SF by fatty acid composition and compare the results with the obtained when using the traditional approach.

Universitat de Girona. Departament d’Informàtica i Matemàtica Aplicada