Lactation and test day random regression models for genetic evaluation of Murrah buffaloes

Abstract

In this study the variance components, genetic parameters and breeding values for 305 day or less milk yield were estimated using lactation model and random regression models for the first three lactations in Murrah buffaloes. Random regression model were taken both as random regression model with homogeneous residual variance (RRMHOM) and heterogeneous residual variance (RRM-HET). The estimates of additive genetic variance using RRM were higher than lactation model in all the three lactations. RRM-HET gave higher estimates of additive genetic variance than RRM-HOM in first lactation while RRM-HOM gave higher estimates in second and third lactation. From RRM, it was possible to account for permanent environmental variance arising due to individual milk yield variations during lactation. The heritability estimates were comparable in all the three models. However, in first lactation, the heritability estimates from lactation model RRM-HOM and RRM-HET were 0.319, 0.296 and 0.305, respectively. Likewise in second and third lactations these estimates were 0.004, 0.137, 0.135 and 0.520, 0.315, 0.264, respectively. The breeding value rank correlation was high in all the lactations. More sires were common for each model among the top 10 ranked sires in all three lactations. In conclusion, RRM can be an alternative to lactation model owing to high accuracy, early evaluation, high additive genetic variance, comparable heritability and high rank correlation for breeding values.