Sheep & Goats

Putting flocks on the scales: evaluating flock performance for lamb liveweight using random regression models

---

Abstract

Random regression models are widely used to model liveweight trajectories across livestock species, enabling the estimation of genetic (co)variance components across age when multiple records are available per animal. Although previous studies have reported limited improvements in the accuracy of genetic evaluations for liveweight using random regression models compared with multi-trait models, these models may offer the potential to derive flock-level random regressions for liveweight that could serve as a useful management tool. The objective of this study was to quantify the contribution of both genetic and non-genetic factors to the variability in lamb liveweight trajectories across age, with a focus on generating flock-specific profiles to facilitate benchmarking across flocks. Liveweight observations from 33,561 lambs weighed between 0 to 150 days of age were available from 461 flocks across 11 years; at least 4 liveweight records were available per animal. Direct genetic, direct permanent environmental, maternal genetic and a maternal litter effect were fitted as random regressions using Legendre polynomials across age in days. A random regression on flock-year across age was included along with a random effect of flock week of weighing to account for different management groups within a flock. The most parsimonious order of the random regression was chosen by comparing the Akaike information criterion. Flock-year trajectories, representing the flock-year performance independent of all other fitted effects, were subsequently analysed by calculating percentile ranks within each year at 40, 100 and 150 days of age. A random regression model with quadratic polynomials for the random effects provided a better fit than lower-order polynomials; higher-order polynomials were tested but failed to converge. Direct additive genetic standard deviation increased across the age trajectory and ranged from 0.28 kg at birth to 2.57 kg at 150 days. Direct heritability varied across the trajectory, rising from 0.06 at birth to a maximum of 0.10 at 76 days of age. Maternal heritability ranged from 0.08 at 11 days of age to 0.03 at 150 days. The proportion of total phenotypic variance attributable to flock-year increased with age, from 0.04 at birth to 0.60 at 150 days. Flock profiles were moderately consistent across years; for flocks with records spanning multiple years, the proportion of total variance in the predicted flock-year profiles explained by differences between flocks increased with age, from 0.48 at day 40 to 0.65 at day 150. Among flocks with records in consecutive years, a flock in the top 20% at day 100 in a given year had a 47% probability of remaining in the top 20% at day 100 in the following year. Results from this study demonstrate that random regression models can generate meaningful flock-level profiles for liveweight, providing a valuable tool for monitoring and benchmarking performance across flocks.

Keywords: 2026

How to Cite:

Dunne, A., Berry, D., Purfield, D., Pabiou, T. & McHugh, N., (2026) “Putting flocks on the scales: evaluating flock performance for lamb liveweight using random regression models”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2275152. doi: https://doi.org/10.31274/wcgalp.23423

Rights: 1