An Approach to Connect Multi-trait Mixed Model and Principal Component Analysis for Describing Variation in Carcass Quality of Crossbred Cattle

A principal component analysis of the 4×4 sire, maternal, management and environmental (co)variance matrices derived from a multi-trait sire model was conducted to describe variability in four economically important carcass traits. Carcass weight (HCWt), P8 fat (P8), eye muscle area (EMA) and intramuscular fat (IMF) collected from 1144 heifers and steers calves from seven sire breeds: Angus, Belgian Blue, Hereford, Jersey, Limousin, South Devon and Wagyu, born over a 4year period. The first two principal components (PC1, PC2) accounted for 90% of the total variance in the considered variables, except for the maternal component, where PC1 and PC2 accounted for 83% of the total variance. The largest and the least variations attributed to the management (99%) and maternal (83%) components, respectively. Sire and environment components showed similar patterns of eigenvector coefficients for the first two vectors. The first and second eigenvectors have large loadings for P8 fat and IMF, respectively. The third orthogonal vector had a large coefficient for the HCWt and EMA but not other traits. For the maternal component, which is a small component of overall variation, P8 fat in contrast to IMF had a significant relationship with the PC1. PC1 could be defined as a fat distribution component. PC2 respects mean values for carcass traits with less attention to EMA, presenting market suitability. For management component as the largest component of overall variation, PC1 could be interpreted as a weighted mean with much more emphasis on the IMF. PC2 accounting, for 25.78% of the total variance, indicated a major contrast between P8 and IMF, consequently it can be interpreted as a fat distribution component.