Development of a multivariate yield grade equation to predict compositional traits in mature cow carcasses

Abstract

Data from 68 cow carcasses were used to develop a new yield grading system. First principal component (FPC) values for compositional attributes (LNFT = separable lean weight/[lean + fat weightl × 100, LNBN = separable lean weight/[lean + bone + connective tissue weightsl × 100, and BTPR = [defatted lean from the round, loin, rib, and chuckl/side weight × 100) were determined. The first component explained 83.5% of the standardized variance and load values were .63, −.52, and .58, respectively. The resulting FPC values ranged from −1.93 to 1.89. The linear regression of LNFT, LNBN, and BTPR (dependent variables) on FPC (independent variable) explained a significant amount of variation (P < .001) in each case and resulted in R²-values of .98, .67, and .85, respectively. A best-fit yield grade equation, developed to predict FPC, included adjusted fat thickness (ADF), percentage of kidney, pelvic, and heart fat (KPH), and overall muscling grade (OM). The equation, FPC = 2.04 − (.67 × ADF) − (.21 × KPH) − (.0016 × OM), explained a significant amount (P < .001) of variation in FPC with R² = .94 and residual standard deviation = .25. Simple correlations for ADF, KPH, and OM with FPC were −.87, −.71, and −.80, respectively. Cow carcasses were assigned to one of three grades based on FPC values that corresponded with predetermined levels of LNFT, LNBN, and BTPR. These grades generally had smaller CV than existing grades. When used in conjunction with quality grades, proposed grades could be more useful to the cow meat industry.