BeefSpecs fat calculator to assist decision making to increase compliance rates with beef carcass specifications: evaluation of inputs and outputs
M. J. McPhee A B D , B. J. Walmsley A B , D. G. Mayer A C and V. H. Oddy A BA Cooperative Research Centre for Beef Genetic Technologies.
B NSW Department of Primary Industries, Beef Industry Centre of Excellence, Trevenna Road, Armidale, NSW 2351, Australia.
C Department of Agriculture, Fisheries and Forestry, Ecosciences Precinct, 41 Boggo Road, Dutton Park, Qld 4102, Australia.
D Corresponding author. Email: malcolm.mcphee@dpi.nsw.gov.au
Animal Production Science 54(12) 2011-2017 https://doi.org/10.1071/AN14614
Submitted: 31 May 2014 Accepted: 22 July 2014 Published: 3 September 2014
Abstract
This study evaluated the BeefSpecs fat calculator, a decision-support system developed to assist the beef industry to increase compliance rates with carcass specifications (weight and fat specifications). A challenge to the BeefSpecs calculator and a sensitivity analysis were used to evaluate the inputs and outputs of BeefSpecs. Five industry datasets (n = 80, 97, 68, 25, and 13 for Datasets 1–5, respectively) of Bos taurus, Bos indicus, and Bos taurus × Bos indicus breeds for steers and heifers were collated to challenge BeefSpecs, and a nine-way factorial matrix (n = 57 600) of input variables was created for the sensitivity analysis. There were no significant (P > 0.05) differences in the mean bias between observed and predicted values in any of the datasets but there were significant (P < 0.01) differences in the unity of slope for Datasets 2, 3, and 5. The root-mean-square error was 1.72, 2.61, 2.87, 2.68, and 2.00 mm for Datasets 1–5. The decomposition of the mean-square error of prediction indicated that most of the error contained in the predictions of all models was of a random nature (94%, 85%, 85%, 95% for Datasets 1–4), except in Dataset 5, which had a 47% proportion of error in the slope component. All datasets indicated little bias (0.13%, 12.19%, 12.69%, 0.60%, and 0.12% for Datasets 1–5) in the model predictions. An analysis of variance with the nine-way factorial matrix on the predicted output of final P8 fat was conducted for the sensitivity analysis. A significant (P < 0.01) four-way interaction of days on feed × frame score × initial liveweight × sex was detected. Final P8 fat was sensitive to measurement error in the inputs of frame score when animals had longer feeding periods (e.g. 180 days) and to initial P8 fat when animals had lower initial liveweights (e.g. 200 kg) and higher frame scores (e.g. 7). For each unit of error in estimating frame score, BeefSpecs predicts final P8 with an error of up to 2.3 mm in heifers and up to 1.7 mm in steers. Error in the estimation of initial P8 fat of 2 mm will result in an error of up to 3 mm in the prediction of final P8 fat. The sensitivity analysis of BeefSpecs input variables (frame score and initial P8 fat) on the prediction of final P8 fat indicates that increasing the accuracy of estimating frame score and P8 fat is an issue that needs addressing.
Additional keywords: cattle, DSS, frame score, P8 fat depth.
References
Bibby J, Toutenburg H (1977) ‘Prediction and improved estimation in linear models.’ (John Wiley & Sons and Akademie-Verlag: Berlin)Dicker RW, Ayres JF, McPhee MJ, Robinson DL, Turner AD, Wolcott ML, Kamphorst PG, Harden S, Oddy VH (2001) Post-weaning growth of cattle in northern New South Wales. 2. Growth pathways of steers. Australian Journal of Experimental Agriculture 41, 971–979.
| Post-weaning growth of cattle in northern New South Wales. 2. Growth pathways of steers.Crossref | GoogleScholarGoogle Scholar |
Keele JW, Williams CB, Bennett GL (1992) A computer model to predict the effects of level of nutrition on composition of empty body gain in beef cattle: I. Theory and development. Journal of Animal Science 70, 841–857.
McKiernan WA, Wilkins JF, Irwin J, Orchard B, Barwick SA (2009) Performance of steer progeny of sires differing in genetic potential for fatness and meat yield following postweaning growth at different rates. 2. Carcass traits. Animal Production Science 49, 525–534.
| Performance of steer progeny of sires differing in genetic potential for fatness and meat yield following postweaning growth at different rates. 2. Carcass traits.Crossref | GoogleScholarGoogle Scholar |
McPhee MJ, Harden S, Robinson DL, Dicker RW, Oddy VH (2012) Effects of backgrounding and finishing growth rates on P8 fat and intramuscular fat in Bos taurus steers. Animal Production Science 52, 354–364.
| Effects of backgrounding and finishing growth rates on P8 fat and intramuscular fat in Bos taurus steers.Crossref | GoogleScholarGoogle Scholar |
R Development Core Team (2009) ‘R: A language and environment for statistical computing.’ (R Foundation for Statistical Computing: Vienna) Available at http://www.R-project.org [Accessed 4 August 2009]
Robinson DL, Oddy VH, Dicker RW, McPhee MJ (2001) Post-weaning growth of cattle in northern New South Wales. 3. Carry-over effects on finishing, carcass characteristics and intramuscular fat. Australian Journal of Experimental Agriculture 41, 1041–1050.
| Post-weaning growth of cattle in northern New South Wales. 3. Carry-over effects on finishing, carcass characteristics and intramuscular fat.Crossref | GoogleScholarGoogle Scholar |
Saltelli A, Annoni P (2010) How to avoid a perfunctory sensitivity analysis. Environmental Modelling & Software 25, 1508–1517.
| How to avoid a perfunctory sensitivity analysis.Crossref | GoogleScholarGoogle Scholar |
Saltelli A, Chan K, Scott PM (2000) ‘Sensitivity analysis.’ Wiley Series in Probability and Statistics. (Wiley: New York)
Upton WH, Donoghue KA, Graser HU, Johnston DJ (1999) Ultrasound proficiency testing. In ‘Proceedings of the Association for the Advancement of Animal Breeding and Genetics. Vol. 3’. pp. 341–344. (Association for the Advancement of Animal Breeding and Genetics)
VSN International (2011) ‘Genstat for Windows, Release 14.1.’ (VSN International Ltd: Hemel Hempstead, UK)
Walmsley BJ, Oddy VH, McPhee MJ, Mayer DG, McKiernan WA (2011) Development of the BeefSpecs fat calculator: a tool designed to assist decision making to increase on-farm and feedlot profitability. In ‘MODSIM 2011. Proceedings 19th international congress on modelling and simulation’. (Eds F Chan, D Marinova, RS Anderssen) pp. 898–904. (Modelling and Simulation Society of Australia and New Zealand)
Walmsley BJ, Mayer DG, McPhee MJ (2013) What impact do producer measured inputs have on the prediction accuracy of BeefSpecs? In ‘MODSIM 2013. Proceedings 20th international congress on modelling and simulation’. (Eds J Piantadosi, RS Anderssen, J Boland) pp. 649–655. (Modelling and Simulation Society of Australia and New Zealand)
Walmsley BJ, McPhee MJ, Oddy VH (2014) Development of the BeefSpecs fat calculator to assist decision making to increase compliance rates with beef carcass specifications. Animal Production Science 54, 2003–2010.
| Development of the BeefSpecs fat calculator to assist decision making to increase compliance rates with beef carcass specifications.Crossref | GoogleScholarGoogle Scholar |
Williams CB, Jenkins TG (1997) Predicting empty body composition and composition of empty body weight changes in mature cattle. Agricultural Systems 53, 1–25.
| Predicting empty body composition and composition of empty body weight changes in mature cattle.Crossref | GoogleScholarGoogle Scholar |
Williams CB, Jenkins TG (1998) A computer model to predict composition of empty body weight changes in cattle at all stages of maturity. Journal of Animal Science 76, 980–987.
Williams CB, Keele JW, Bennett GL (1992) A computer model to predict the effects of level of nutrition on composition of empty body gain in beef cattle: II. Evaluation of the model. Journal of Animal Science 70, 858–866.
Williams CB, Bennett GL, Keele JW (1995) Simulated influence of postweaning production system on performance of different biological types of cattle: I. Estimation of model parameters. Journal of Animal Science 73, 665–673.
