130 Bain et al.- Prediction of lean meat yield BRIEF COMMUNICATION: Carcass linear measurements as predictors of meat yield in lambs determined by VIAscan W.E. BAIN*, J.C. MCEWAN, N.J. MCLEAN, and P.L. JOHNSON AgResearch Invermay, Private Bag 50-034, Mosgiel 9053, New Zealand *Corresponding author: wendy.bain@agresearch.co.nz Keywords: lamb; carcass; lean meat yield; VIAscan. INTRODUCTION Lambs yielding proportionally more lean meat are desirable to processing companies as they have a greater quantity of saleable product which also meets market demand without the need for further processing. The quest for higher yielding lean lambs has seen a change in the lamb carcass grading system from classification on carcass weight and the GR measurement (total tissue depth above the 12th rib, 110 mm from the midline) (Kirton, 1989) to using imaging systems such as VIAscan®. VIAscan® is a two dimensional imaging system that estimates lean content of the carcass (Hopkins et al., 2004). Hopkins et al. (2004) have shown that using VIAscan® versus the traditional carcass weight and GR measurements increases the prediction accuracy of estimating the lean meat yield of a carcass (R 2 = 0.52 versus R 2 = 0.19). The introduction of VIAscan® in some meat plants has enabled online predictions of the proportion of lean meat in the total carcass as well as the leg, loin and shoulder regions. This technology allows meat companies to reward producers for meat yield within a carcass region, an incentive for farmers to produce higher lean meat yielding lambs (Jopson et al., 2009). A data set was collected to search for genetic markers for lean meat yield, using high and low yielding carcasses, based on VIAscan® estimates of total yield. This provided an opportunity to also investigate relationships between linear carcass measurements and lean meat yield as assessed by the VIAscan® grading. This work is the focus of this report. A transition from G to A in the 3’ untranslated region of the GDF8 gene (c.1232 G>A; (Hickford et al., 2009)), derived from Texels, is associated with increased lean meat yield and is also associated with increased buttock circumference (Johnson et al., 2009). The relationship between increased lean meat yield and butt circumference both with and without adjustment for this mutation is also investigated. MATERIALS AND METHODS Data collection was carried out between January and April in 2008 and 2009. Mobs of lambs were observed at Alliance Mataura meat processing plant as they travelled through the VIAscan®. To reduce variation in the results, criteria included carcasses from ram lambs only, selected from large mobs of greater than 200 lambs, with a carcass weight between 15.5 and 19.0 kg. One to three most extreme yielding pairs with a high and a low carcass yield, matched for carcass weight, were identified from the selected mobs. This Brief Communication looks at a subset of 834 carcasses, from the 2008 data set which had been genotyped for the GDF8 c.1232 G>A mutation and represented 209 different mobs. Measurements recorded on the whole carcass were cold carcass weight (CW), GR depth and carcass linear measurements of buttocks circumference (BC), carcass length (CL), and leg length (LL). VIAscan® carcass measurements of leg yield, loin yield, shoulder yield, and total yield were recorded and expressed as a percentage of the carcass weight. Carcass length was measured from between the hind legs to the front of the neck using a set of callipers with 50 mm wide bars at each end. Leg length was measured from the crotch to the end of the hind leg, which was cut though the tarsal joint. The circumference of the buttocks was measured using a flexible tape measure on the dressed carcasses hanging from their hindquarters and represented the circumference when taken in a parallel plane immediately above the anal opening. A meat sample was collected from all lambs measured for DNA. These meat samples were genotyped for the GDF8 c.1232 G>A mutation. The A allele is associated with the increased lean meat yield phenotype. The protocol of Johnson et al. (2009) was used, with 611, 194, 29 carcasses carrying 0 (GG), 1 (AG) or 2 (AA) copies of the A allele respectively. Data were analysed using a stepwise regression procedure in SAS (SAS, 2004) to determine which measurements best predicted VIAScan® estimated yield. The three carcass regions of leg yield, loin yield, shoulder yield and total yield were the dependant variables, within each of the A allele groups (GG, AG or AA). Variables, BC, GR, CW, LL, and CL were included in the stepwise analysis. RESULTS AND DISCUSSION Summary statistics for the measurements made in the meat processing plant are given in Table 1.