45167

Genetic heterogeneity of residual variance of yearling weight evaluated under different approaches in Nellore cattle.

Favorite this paper

Evidences of genetic heterogeneity of residual variance have been found on studies involving several species, traits, and approaches. In beef cattle, there is great interest in increasing uniformity of the final product, which could be achieved through selection for this putative genetic component of residual variance. The aim of this study was to investigate the possibility to reduce the variability around the mean of yearling weight (YW, measured at 550 days) in a Nellore beef cattle population, using two statistical approaches. For this, records of YW from 194,628 animals from Alian?a Nelore database were used. Only sires with a minimum of 50 progenies and contemporary groups (CG) with at least 25 animals were used. In the statistical analysis, a sire model including fixed effect of CG (herd and season of birth, management group at weaning and yearling, and sex) and age at measurement (linear and quadratic effects) as a covariate, was fitted under two-step (Mulder et al., 2009) and double hierarchical generalized linear model (DHGLM, Felleki et al., 2012) using the DMU package (Madsen and Jensen, 2013). In the first step of the two-step approach, models assuming homogeneous and heterogeneous residual variances for each sire were adopted. Heritability () and genetic coefficient of variation (GCVE) of residual variance were calculated, aiming to compare such estimates with previous studies as well as to guide the selection for uniformity. Lower estimates of genetic parameters were obtained compared to literature. Estimates of (about 0.005 and 0.012 for two-step and DHGLM, respectively) obtained highlight the need for large sire families to predict breeding values for residual variance accurately. The opportunity to improve uniformity through genetic heterogeneity of residual variance, measured by GCVE, was lower under two-step approach (11%) than DHGLM (17%). The positive correlation between genetic effect on the mean and residual variance (about 0.20 and 0.76 for two-step and DHGLM, respectively) is unfavorable, if the aim is increasing mean and reducing variability, although both criteria can be included in a selection index. The findings of both approaches suggest the existence of a genetic component on residual variance of YW in Nellore cattle.