Genetic parameter estimates for growth traits in Katahdin sheep using different models

Genetic parameters for growth traits of Katahdin lambs were estimated using six variants of the animal model. Data on birth weight (BW; n= 13,099), weaning weight adjusted to 75 d (WW; n=11,509) and postweaning weight adjusted to 120 d (AW; n=6,886) were collected for seven years (2004-2010) in 20 states across Mexico. Analyses were carried out by ignoring or including maternal effects. The simplest model included the direct additive genetic effect as the only random effect. The most complete model included direct and maternal genetic effects, their covariance, and the maternal permanent environmental effect. Selection of the best model was based on likelihood-ratio test. When maternal effects were not taken into account, estimates of direct heritability and direct genetic variance were overestimated for all traits. Direct heritability estimates for the best model were 0.18 +/- 0.03, 0.30 +/- 0.04, and 0.20 +/- 0.05 for BW, WW and AW, respectively. Maternal heritability estimates also varied depending on the model; estimates ranged from 0.05 to 0.23, 0.00 to 0.12, and 0.09 to 0.25 for BW, WW and AW. Ignoring maternal effects in the model would result in inaccurate genetic evaluation for growth traits in Katahdin sheep.