CORRELATIONS OF CARCASS AND MEAT QUALITY TRAITS IN CROSSBRED AND PURE BULLS FEEDLOT FINISHED THROUGH MULTIVARIATE APPROACH

Multivariate approaches allow identifying the most important directions of variation, classifying meat samples into subgroups of similar behavior. Principal component analysis (PCA) and the clustering analysis (K-means) were used to characterize carcass and meat quality traits of F1 Angus-Nellore and Nellore bulls feedlot finished for. The variables comprised initial body weight(BW), final BW at slaughter, hot carcass weight, carcass yield, backfat thickness, rib eye area, pH (48h postmortem), cooking losses (CL), meat colour (lightness, L*; redness, a*; yellowness, b*; chromaticity [Chroma] and real color [Hue]), Warner–Bratzler shear force (WBSF), collagen, intramuscular fat (IMF), protein, moisture and ash. The PCA was performed to characterize variables in carcass and longissimus muscle. Considering WBSF (tenderness) as a separation criterion, the K-means (k = 2) cluster analysis was applied to classify animals in two groups. The five first principal component (PC) explained about 71.7% of total data variability and meat toughness (i.e. positive coefficients for WBSF and CL) was more effective to define the PC1. The meat color variables (L*, a*, b* Chroma, and Hue) and IMF plus moisture were more effective to define the second (PC2) and third component (PC3), respectively. In this study, beef samples with higher values of WBSF (tough meat) commonly showed lower chromaticity, yellowness and redness values. The projection of WBSF groups allowed to distinguish the two genotypes groups (F1 Angus-Nellore versus Nellore bulls) in relation to PC1 (meat toughness plus high losses) and PC2 (meat colour aspect), as well as in PC1 and PC3 (marbling plus low moisture).