Identification of relationship between pork colour and physicochemical traits in American Berkshire by canonical correlation analyses

This study was carried out to predict the relationship between the colour and physicochemical traits in pork by using canonical correlation analysis. The variables of pork colour traits were lightness (L*), redness (a*) and yellowness (b*), whereas the variables of physicochemical traits were post-mortem pH24 h, water-holding capacity (WHC), collagen content, fat content, moisture content, protein content, drip loss, cooking loss and shear force. The canonical correlation coefficient (0.819) between the first pair of canonical variates, V1 and W1, was significant (P < 0.01). According to cross loadings, drip loss, cooking loss and fat content provided the relatively high positive correlations with the variates of colour traits (V1), while pH24 h, WHC and moisture content displayed negative relationships with the variates. Otherwise, L* and a* strongly contribute to the variates of physicochemical traits (W1). In addition, a redundancy index of 0.256 suggests that 25.6% of the variance in V1 is explained by W1. Therefore, in order to obtain the reddish pink colour in pork that is preferred by consumers depending on the relationships between the pork quality traits, we suggest that producer leads to the proper maintenance of post-mortem pH24 h, higher WHC, lower drip loss and cooking loss in pork.