Estimates of heritabilities, common litter of birth effects and additive maternal genetic effects were produced for ultrasonic backfat depth, average daily food intake, average daily gain and food conversion ratio of Landrace boars and gilts. Boars and gilts were performance tested under different regimes. A bivariate derivative-free restricted maximum likelihood procedure was used to estimate genetic correlations between the performance test traits as recorded in the two sexes.
Heritability estimates from the analysis including the common litter of birth effect tended to be towards the low end of the range of recently published estimates. This may reflect either population specific effects, such as effects of long-term selection, or the use of an individual animal model.
Estimates of the common litter of birth effect were around 0·05, and generally had a significant effect upon the fit of the model, while additive maternal genetic effect estimates were negligible. Therefore, it is expected that omission of maternal effects from models for evaluation by best linear unbiased prediction will not hinder genetic progress. Inclusion of common litter of birth effects would be recommended, although this result may not hold for populations given food ad libitum.
The estimates of genetic correlations between performance test traits of boars and gilts indicate that the levels of genotype-environment interaction (G × E) and genotype-sex interaction were low across most traits and data sets, with all genetic correlation estimates lying between 0·8 and 1·0. The lowest estimates of the genetic correlations, which were observed in data sets where the environments appeared to differ most, indicate that G × E interactions may be a problem in populations where males and females are subject to test regimes with greater differences than those seen here.