Genetic parameters of weaning-to-oestrus interval in pigs using Bayesian analysis

Abstract

Genetic parameters for weaning-to-oestrus interval (WOI) were estimated on 11,026 sows from pig selection farm Ptuj (Slovenia). Univariate analyses of observations of WOI across parities and two-trait analyses of records of primiparous (PP) and multiparous (MP) sows were performed applying Bayesian approach. WOI was analysed on original scale, transformed using a natural log transformation (lnWOI) and considered as a mixture of normal and exponential distributions (trWOI). Statistical model included genotype, season of service as year-month interaction, and parity as fixed effects, and previous lactation length and number of piglets weaned as covariates. Direct additive genetic effect and permanent environment were considered as random effects. (Co)variance components were estimated by Bayesian analyses. Gibbs chains of length between 130,000 and 180,000 were run. Univariate analysis for sows across parities resulted in heritability estimates of 0.06, 0.07, and 0.07 for WOI, lnWOI and trWOI, respectively. Permanent environment (PE) accounted for 26 % of phenotypic variation for WOI and 19 % for both lnWOI and trWOI. In bivariate analysis, heritability was 0.19 in PP and 0.06 in MP. Genetic correlation for WOI between PP and MP was 0.80. The PE proportion was 0.10 in MP sows. Based on genetic parameters, selection on improved rebreeding performance in pigs is expected to be most efficient considering WOI in primiparous and multiparous sows as different traits. Transformation of WOI increased the genetic parameters only slightly.