J. Anim. Sci. 2003. 81:2103-2104
© 2003 American Society of Animal Science
Consideration of fat thickness in models to predict beef carcass cutability
Michael L. Thonney
Department of Animal Science, 114 Morrison Hall, Cornell University, Ithaca, NY 14853-4801, email: mlt2{at}cornell.edu
Subcutaneous fat thickness is the most logical predictor of wholesale cut yield because it represents what is to be trimmed from the beef carcass. Furthermore, mechanical or ultrasonic fat thickness measurements could be automatically made and recorded at several critical points—not just between the 12th and 13th ribs—to increase the prediction of yield. Adjusted fat thickness (ADJFAT) is representative of such a set of measurements, yet the residual standard deviations (RSD) for predictions based on ADJFAT were not presented by Cannell et al. (2002)
for comparison with other sets of predictors, including those from video images. Because r2 must increase as predictors are added to equations, RSD provide better assessments of the predictive value of equations with varying numbers of predictors.
From the data in Table 5 of Cannell et al. (2002), the total corrected sums of squares were calculated to be 1,580, 2,500, and 437 for percentages of trimmed wholesale cuts, fat trim, and bone in cold carcass, respectively. Based on the respective correlations for ADJFAT of -0.75, 0.87, and -0.46 in Table 2 of Cannell et al. (2002), the RSD were calculated and are presented in Table 1 of this letter for comparison with other predictors from Table 5 of Cannell et al. (2002).
Adjusted fat thickness represents a single predictor, whereas the other combinations of predictors include multiple predictors or single predictors that are composites of multiple predictors. For example, EXPYG represents the expert application of hot carcass weight, longissimus muscle area, fat thickness, and kidney, pelvic, and heart fat (KPH) in the current USDA yield grade equation, which was developed from data from small-framed cattle (Murphey et al., 1960). Adjusted fat thickness predicted trimmed wholesale cuts with a RSD only 0.23 percentage units higher than the more time-consuming expert yield grade predictor and only 0.17 percentage units higher than when hot carcass weight and cold camera longissimus muscle area measurements were added as predictors. Adjusted fat thickness predicted trimmed wholesale cuts with 0.25 percentage units lower RSD than the online grader yield grade it could replace immediately. Moreover, prediction of fat trim and waste bone by ADJFAT resulted in RSD essentially as low as any other set of predictors (Table 1
). Cannell et al. (2002) indicated that ADJFAT had the single highest correlation with fabricated cuts. Given the ease with which fat thickness data collection could be automated and the precision with which it estimates fabrication waste and yield, it is surprising that they did not include it as a single-predictor regression model to compare with their other models in Table 5.
The underlying biology of growth and carcass composition provides a more compelling reason than the empirical data of Cannell et al. (2002) to choose simple fat thickness measurements to predict carcass waste and yield: Compared with other predictors, fat thickness is unbiased. Because larger mature-sized animals gain less fat and more muscle to a given carcass weight than smaller mature-sized animals (Berg and Butterfield, 1976; Nour et al., 1981, 1983), use of hot carcass weight as a predictor is biased. Because longissimus muscle area provides only a two-dimensional measurement of a three-dimensional object, use of longissimus muscle area as a predictor is biased against cattle with longer muscles (Nour et al., 1981). Furthermore, in a commercial setting, KPH cannot be measured objectively, and significant increases in efficiency could be made if KPH were removed in the processing plant, as is the case in Canada. Because it is an unbiased and easily measured predictor, multiple mechanical or ultrasonic probe measurements of subcutaneous fat thickness should be used to predict carcass fabrication waste and yield.
Literature Cited
Berg, R. T., and R. M. Butterfield. 1976. New Concepts of Cattle Growth. John Wiley & Sons, New York.
Cannell, R. C., K. E. Belk, J. D. Tatum, J. W. Wise, P. L. Chapman, J. A. Scanga, and G. C. Smith. 2002. Online evaluation of a commercial video image analysis system (Computer Vision System) to predict beef carcass red meat yield and for augmenting the assignment of USDA yield grades. J. Anim. Sci. 80:1195–1201.[Abstract/Free Full Text]
Murphey, C. E., D. K. Hallett, W. E. Tyler, and J. C. Pierce Jr. 1960. Estimating yields of retail cuts from beef carcasses. J. Anim. Sci. 19:1240.
Nour, A. Y. M., M. L. Thonney, J. R. Stouffer, and W. R. C. White, Jr. 1981. Muscle, fat and bone in serially slaughtered large dairy or small beef cattle fed corn or corn silage diets in one of two locations. J. Anim. Sci. 52:512–521.[Abstract/Free Full Text]
Nour, A. Y. M., M. L. Thonney, J. R. Stouffer, and W. R. C. White Jr. 1983. Changes in primal cut yield with increasing weight of large and small cattle. J. Anim. Sci. 57:1166–1172.[Abstract/Free Full Text]
