Random regression test-day milk yield models as a suitable alternative to the traditional 305-day lactation model for genetic evaluation of Sahiwal cattle

Abstract

In the present study, first three lactation 305-day milk yield variance components, genetic parameters and breeding value were estimated using test-day milk yield and 305-day actual milk yield data of Sahiwal cattle. The estimates were obtained using three methods viz. random regression model (RRM) with homogeneous residual variance (RRM-HOM), RRM with heterogeneous residual variance (RRM-HET) and univariate animal model. The additive genetic variance of 305-day milk yield estimated from RRM was higher compared to univariate animal model for all lactation. From RRM, it was possible to account for permanent environmental effects due to individual milk yield variations during lactation. The heritability estimates were low for first (0.072 to 0.079) and third lactation (0.087 to 0.112) 305-day milk yield from all three methods. For second lactation, low heritability estimate from univariate animal model (0.144) and moderate estimate from different RRM (0.206 to 0.219) were obtained. For all lactation, breeding value rank correlation was more than 0.78 between lactation model and random regression testday model. The same bull was identified as top ranking bull from all three methods. It can be concluded that random regression test-day models can replace conventional 305-day lactation model for genetic evaluation as it resulted in higher additive genetic variance estimates, gave similar/higher heritability value and moderate to high rank correlation estimates for breeding values.