Estimation of the Variance Components of the Sow Litter Size Traits Using Reml Method - Repeatability Model

Abstract

Variance components for sow litter size traits were estimated using the REML method. Number of live born piglets (NBA), number of still born piglets (NSB), number of total born piglets (NTB) and number of weaned piglets (NW) were treated as traits which repeated several times during sow lifetime - repeatability model. Results of the fertility of Swedish Landrace sows realized on three pig farms in the Republic of Serbia were presented in four data sets DS1 (farm 1), DS2 (farm 2), DS3 (farm 3) and DS23 (farms 2 and 3 together). Fixed part of the model for litter size traits at farrowing (NBA, NSB and NTB) included parity, mating season as year-month interaction, litter genotype and weaning to conception interval as class effects. The age at farrowing was modelled as a quadratic regression nested within parity, whereas preceding lactation length was included as linear regression. In case of NW the model included parity, weaning season as year-month interaction, number of piglets in litter subsequent to crossfostering and litter genotype as class effects. The age at farrowing was included into the model in the same way as in case of previous traits. Random part of the model was the same for all analysed traits and represented as effect of common environment in litter where sows had been born, permanent effect of environment in sows’ litters and direct additive genetic effect. Heritability of NBA varied between 0.050 (DS2) and 0.076 (DS3), NSB between 0.004 (DS3) and 0.027 (DS2), NTB between 0.065 (DS2) and 0.073 (DS3) and of NW between 0.010 (DS2) and 0.028 (DS1). Share of permanent environment of sow in phenotypic variance was higher than share of litter effect and mostly lower than share of direct genetic effect.