Ham- and shoulder-face characteristics in relation to pork primal composition

The faces of hams (n=249) and shoulders (n=248) separated from the middles of hog carcasses according to standard Canadian breaking lines, were imaged with a consumer camera. Primals were then scanned with a GE Lunar iDXA unit using dual energy X-ray absorptiometry (DXA), and fat and lean compositions were calculated with pre-established high accuracy prediction equations. Mean ham and shoulder % fat compositions were 21.29 (SD = 2.75; range = 14.27 – 29.22) and 26.48 (SD = 3.60; range = 17.66 – 35.96) respectively. Face images underwent bioimage analysis, determining proportions and thicknesses of fat and lean. Stepwise regression of image measurements on DXA results showed that fat area and fat thicknesses of the ham face were the measurements best suited for estimation of % fat and lean obtained from DXA (model R2 = 0.72 and 0.73 respectively; Pr > F: < 0.0001), while addition of primal weight for estimation of grams of fat and lean gave model R2 of 0.76 and 0.93 respectively (Pr > F: < 0.0001). For the shoulder, a combination of two fat area measurements, and fat and lean depths were found to best estimate % fat and lean obtained from DXA (model R2 = 0.70 and 0.69 respectively; Pr > F: < 0.0001), while addition of primal weight for estimated grams of fat and lean gave model R2 of 0.80 and 0.85 respectively (Pr > F: < 0.0001). Implications: Dimensional measurements of ham or shoulder face characteristics may enable estimation of primal composition very early on the carcass fabrication line.