Estimation of autumn live weight breeding values in progeny test datasets when progeny are slaughtered to achieve a target carcass weight

N.B. JOPSON, J.C. MCEWAN, K.J. KNOWLER and R. WHEELER

Proceedings of the New Zealand Society of Animal Production 67: 175-179

An analysis was undertaken to investigate how lamb autumn live weight breeding values could be calculated from fixed carcass weight progeny tests (e.g. lambs slaughtered at monthly intervals at a threshold carcass weight). Analysis of growth rate is problematic for this progeny test design as the slaughter groups may experience vastly different environmental conditions, both within and between years.

Variance components were estimated from an animal model using restricted maximum likelihood (REML) on data from 11,411 Coopworth progeny records and 126 Coopworth sires. The dataset contained records of weaning weight and live weight at four, five and six months of age (LW4, LW5 and LW6, respectively) collected from 1995 to 2004 in a flock run at Woodlands Research Station. The multivariate mixed linear model included lamb birth day deviation as a covariate, sex of lamb, birth rearing rank, age of dam and live weight contemporary group as fixed effects, and sire as a random effect. The WWT model also had dam fitted as a random effect to give WWT direct and maternal effects (WWTd and WWTm). The analysis was then repeated, but with animals records removed from the dataset once they achieved a threshold weight of 31 and 27kg for ram and ewe lambs, respectively, on data prior to or following linear scaling by contemporary group mean.

The heritabilities estimated from the full dataset for WWTd, WWTm, LW4, LW5 and LW6 were 0.30±0.03, 0.04±0.01, 0.49±0.02, 0.49±0.02 and 0.51±0.02, respectively. These traits were highly correlated with both phenotypic and genetic correlations between the direct traits ranging from 84 to 99%. The same analysis on the two culled datasets resulted in only minor changes the variance components. The correlation between sire breeding values for autumn live weight (LW6) estimated from the full and the two culled datasets were 98.8% for both models. Autumn live weight breeding values can be accurately estimated from datasets which include progressive culling if a multivariate live weight analysis is undertaken, and data from all animals measured at each date are used.

Keywords: NZSAPAB; Sheep; genetic parameters; progeny test

Last Updated 18/07/2007