Quanti fi cation of Saleable Meat Yield Using Objective Measurements Captured

ABSTRACT: Video image analysis (VIA) images from grain-fi nished beef carcasses [n = 211; of which 63 did not receive zilpaterol hydrochloride (ZIL) and 148 received ZIL before harvest] were analyzed for indicators of muscle and fat to illustrate the ability to improve methodology to predict saleable meat yield of cattle fed and not fed ZIL. Carcasses were processed in large commercial beef processing facilities and were fabricated into standard subprimals, fat, and bone. Images taken by VIA technology were evaluated using computer image analysis software to quantify fat and lean parameters which were subsequently used in multiple-linear regression models to predict percentage of saleable meat yield for each carcass. Prediction models included variables currently quantifi ed by VIA technology such as LM area (LMA), subcutaneous (SC) fat thickness at 75% the length of the LM (SFT75), and intramuscular fat score (IMF). Additional distance and area measures included LM width (LW), LM depth (LD), iliocostalis muscle area (IA), SC fat thickness at 25, 50, and 100% the length of the LM (SFT25, SFT50, SFT100), SC fat area from 25 to 100% the length of the LM (SCFA), and SC fat area adjacent to the 75% length of the LM from the spinous processes (SCFA75). Multiple ratio and product variables were also created from distance and area measures. For carcasses in this investigation, a 6 variable equation (Adj. R2 = 0.62, MSE = 0.022) was calculated which included coeffi cients for ZIL treatment, SCFA75, LW, SCFA, SCFA/HCW, and SFT100/HCW. Use of parameters in the U.S. (Adj. R2 = 0.39, MSE = 0.028) and Canadian [Adj. R2 = 0.10, root mean square error (MSE) = 0.034] yield grade equations lack the predictability of the newly adapted equations developed for ZIL-fed and non-ZIL-fed cattle. Prediction equations developed in this study indicate that the use of VIA technology to quantify measurements taken at the 12th/13th rib separation could be used to predict saleable meat yield more accurately than those currently in use by U.S. and Canadian grading systems. Improvement in saleable meat yield prediction has the potential to decrease boxed beef variability via more homogeneous classifi cation of carcass fabrication yield.