Prediction of sheep carcass traits from early-life records using machine learning

Currently hot carcass weight (HCW) and fat score jointly indicate the price grid for sheep meat in Australia. However, experts in the field believe that soon, yield and quality traits such as intramuscular fat (IMF), greville rule fat depth (GRFAT), computed tomography lean meat yield (CTLEAN), and loin weight (LW) are likely to play a role in pricing. Having an accurate prediction of these traits earlier in the life of an animal will allow sheep producers to adjust their management practices in order to achieve the target market requirements. Management, genetics, pasture and climate factors, influence these traits directly and epistatically. Traditional prediction methods may not be powerful enough to capture complex interactions while avoiding overfitting. In this case, learning algorithms that can learn from the current data to predict the animal's future performance offers promise. In this study, five different types of Machine Learning (ML) algorithm, namely Deep Learning (DL), Gradient Boosting Tree (GBT), K-Nearest Neighbour (KNN), Model Tree (MT), and Random Forest (RF) were employed to predict HCW, IMF, GRFAT, LW and CTLEAN and their performances were compared against linear regression (LR) as the gold standard of multinomial prediction. Four scenarios representing different numbers of weight recordings-from a total of 9 weight measures taken between birth (WT1) and pre-slaughter (WT9)- were used to inform the algorithms and all models were trained and tested under equal conditions with identical training and testing sets. Selection of the most effective subset of predictor features were completed via greedy stepwise search among all the available features jointly with expert opinion. In predicting all the traits, RF was superior while LR and KNN showed the lowest prediction performance. When using the final model for predicting on an independent test set, the scenario with the most accurate prediction performance differed across traits. IMF and GRFAT were most accurately predicted when using birth, weaning, and pre-slaughter weights, while the most accurate scenario for HCW, LW and CTLEAN utilised weaning, six monthly weight measures after weaning and pre-slaughter weight. Across all scenarios the least accurate prediction was for IMF.

[1]  H. Cross,et al.  Using Ultrasound, Linear Measurements and Live Fat Thickness Estimates to Determine the Carcass Composition of Market Lambs , 1989 .

[2]  K. Stanford,et al.  Methods of predicting lamb carcass composition: A review , 1998 .

[3]  Ian Witten,et al.  Data Mining , 2000 .

[4]  Trevor Hastie,et al.  An Introduction to Statistical Learning , 2013, Springer Texts in Statistics.

[5]  W. M. Robertson,et al.  The prediction of saleable meat yield in lamb carcasses , 1996 .

[6]  S. Shahinfar,et al.  Machine learning approaches for early prediction of adult wool growth and quality in Australian Merino sheep , 2018, Comput. Electron. Agric..

[7]  J. R. Stouffer,et al.  Muscle metabolism and real-time ultrasound measurement of muscle and subcutaneous adipose tissue growth in lambs fed diets containing a beta-agonist. , 1986, Journal of animal science.

[8]  G. Gardner,et al.  Associations of sire estimated breeding values and objective meat quality measurements with sensory scores in Australian lamb. , 2014, Meat science.

[9]  A. D. Mitchell,et al.  Non-invasive methods for the determination of body and carcass composition in livestock: dual-energy X-ray absorptiometry, computed tomography, magnetic resonance imaging and ultrasound: invited review , 2015, Animal : an international journal of animal bioscience.

[10]  A. Kempster,et al.  Prediction of the lean content of lamb carcasses of different breed types , 1976, The Journal of Agricultural Science.

[11]  P. Ślósarz,et al.  Assessment of intramuscular fat content in the longissimus dorsi muscle in live lambs on the basis of ultrasound image , 2001 .

[12]  S. Loerch,et al.  Effects of restricted feeding on growth performance and carcass composition of lambs. , 1994, Journal of animal science.

[13]  T. Mitsuhashi,et al.  Prediction of carcass composition and individual carcass cuts of Japanese Black steers. , 2014, Meat science.

[14]  G. Gardner,et al.  Carcase weight and dressing percentage are increased using Australian Sheep Breeding Values for increased weight and muscling and reduced fat depth. , 2015, Meat science.

[15]  Agnieszka Ludwiczak,et al.  Artificial neural network analysis of ultrasound image for the estimation of intramuscular fat content in lamb muscle , 2011 .

[16]  Andy Liaw,et al.  Classification and Regression by randomForest , 2007 .

[17]  D. Hopkins.  Predicting the weight of lean meat in lamb carcasses and the suitability of this characteristic as a basis for valuing carcasses. , 1994, Meat science.

[18]  Severiano R. Silva,et al.  Use of ultrasonographic examination for in vivo evaluation of body composition and for prediction of carcass quality of sheep , 2017 .

[19]  G E Gardner,et al.  Sire carcass breeding values affect body composition in lambs--2. Effects on fat and bone weight and their distribution within the carcass as measured by computed tomography. , 2015, Meat science.

[20]  G. Smith,et al.  Predicting Carcass Cutability of Rambouillet Rams Using Live Animal Traits , 1977 .

[21]  D W Pethick,et al.  Dual X-ray absorptiometry accurately predicts carcass composition from live sheep and chemical composition of live and dead sheep. , 2009, Meat science.

[22]  J. E. Edwards,et al.  Genetic parameters for meat quality traits of Australian lamb meat. , 2014, Meat science.

[23]  C. D. Page,et al.  Optimization of reproductive management programs using lift chart analysis and cost-sensitive evaluation of classification errors. , 2015, Journal of dairy science.

[24]  Lewis Kahn,et al.  ASKBILL as a web-based program to enhance sheep well-being and productivity , 2017 .

[25]  D. Hopkins,et al.  Predicting the composition of lamb carcases using alternative fat and muscle depth measures. , 2008, Meat science.

[26]  J. Díez,et al.  Using machine learning procedures to ascertain the influence of beef carcass profiles on carcass conformation scores. , 2006, Meat science.

[27]  F Goyache,et al.  Artificial intelligence techniques point out differences in classification performance between light and standard bovine carcasses. , 2003, Meat science.

[28]  Kurt Hornik,et al.  Open-source machine learning: R meets Weka , 2009, Comput. Stat..

[29]  R. G. Banks,et al.  Australian prime lamb—a vision for 2020 , 2006 .

[30]  Rex M. Butterfield,et al.  New Concept of Sheep Growth , 1988 .

[31]  R Core Team,et al.  R: A language and environment for statistical computing. , 2014 .

[32]  K. Kelman.  Lamb growth impacts muscle oxidative capacity and meat quality , 2014 .

[33]  Johnson Dl,et al.  Interrelationships between GR and other lamb carcass fatness measurement , 1979 .

[34]  D. Hopkins,et al.  The prediction of meat yield in lamb carcasses using primal cut weights, carcass measures and the Hennessy Grading Probe , 2011 .