Parent-of-origin effects in sheep: impact on the data fit, explanatory power of the model and theoretical accuracy of the breeding values.

Although parent-of-origin effects have long been recognized, little is known about their association with growth-related traits in sheep. Specially, it is unclear whether their presence in the model can increase the theoretical accuracy of the breeding values or not. In this study, the effect of including parent-of-origin effects in the genetic evaluation model on the data fit, explanatory power of the model and theoretical accuracy of the breeding values was studied for birth weight (BW), weaning weight (WW), 6-months weight (W6) and pre- and post-weaning average daily gains (ADGa and ADGb, respectively) in Moghani sheep. This was done in a two-step process. In the first step each trait was analyzed with a series of 6 animal models which included different combinations of direct and maternal effects. For each trait, the best model (Model I) was selected by Akaike's Information Criterion (AIC). In the second step, by employing 3 additional models, maternal imprinting, paternal imprinting or both of them were added to the model I to determine whether their presence in the model will lead to an improvement in the likelihood of the model or not. The final best model for each trait was selected at this step. The explanatory power of models was measured using the mean squared error of prediction (MSE) and Pearson's correlation coefficient between the observed and predicted values of records (r([Formula: see text],[Formula: see text])). Correlations between traits due to additive and imprinting effects were estimated using bivariate analyses. Except for ADGb, for other traits studied, including imprinting effects in the model improved data fitting evidenced by smaller residual variance, higher r([Formula: see text],[Formula: see text]) and lower MSE. For all traits studied, the significant imprinting effects were of maternal origin (i.e., characterized by silenced maternal alleles/paternal expression). Both the direct and maternal additive genetic variances decreased following presenting maternal imprinting effects to the model, with higher impact on the maternal genetic variance. The latter result showed presence of confounding between direct and maternal additive genetic effects with maternal imprinting effects. The presence of maternal imprinting effects in the model led to an increase in the theoretical accuracy of the additive breeding values from 8% (BW) to 17% (ADGa). Estimates of maternal imprinting heritability ([Formula: see text]), were 0.06 ± 0.03, 0.10 ± 0.01, 0.10 ± 0.05 and 0.17 ± 0.09 for BW, WW, W6 and ADGa, respectively. Direct heritability ranged between 0.02 ± 0.02 (W6) and 0.13 ± 0.04 (ADGa). While maternal imprinting correlations between traits were all positive, ranged from 0.56 ± 0.31 (W6-ADGa) to 0.99 ± 0.18 (WW-W6), direct additive genetic correlations between traits were lower than maternal imprinting correlations and ranged from -0.17 ± 0.06 (BW-ADGb) to 0.94 ± 0.05 (WW-W6). Since inclusion of imprinting effects in the model lead to improve in the explanatory power of the model, significantly better data fit and a significant increase in the theoretical accuracy of additive breeding values, inclusion of imprinting effects in the model for genetic evaluation of Moghani lambs for growth traits was recommended.