CHARACTERIZATION OF BEEF FATTY ACID PROFILE BY MULTIVARIATE ANALYSIS

Eleven parameters (ether extract, EE; SFA; MUFA; PUFA; n-3; n-6; PUFA/SFA, n-6/n-3, h/H ratios; AI, TI indices) obtained from fat content and fatty acids profile of longissimus thoracis muscle of Piemontese (P), Friesian (F) and Limousin (L) breeds were considered. The data were analysed by hierarchical cluster analysis (HCA), principal component analysis (PCA) and canonical discriminant analysis (CDA). HCA produced three clusters. The first cluster was characterized by L animals (6/11), the second by P (8/10) the third by F (6/10); EE, SFA, MUFA contents, PUFA/SFA ratio, AI and h/H indices significantly differed between breeds. PCA showed that PC1 differentiates between fatty and lean meat; F had a higher EE, SFA and MUFA content and also unfavourables AI and TI indices; P beef showed a better PUFA/SFA and h/H ratios and an unfavourable n-6/n-3 ratio. CDA showed that MUFA, EE, SFA and PUFA/SFA, TI index, n-3, AI and h/H ratio were the most discriminating variables. The 80.6% of grouped cases were correctly classified; function 2 was able to distinguish P and L groups; P group has a better PUFA/SFA ratio, while L had a better TI and a higher n-3 content.