Genetic analysis of competition effects for production traits in swine.

摘要

The effects of competition on estimates of genetic parameters for average daily gain (ADG) during test, days to 110 kg (D110), daily feed intake (DFI), p2 backfat (BF), and muscle depth (MD) were studied. Records from 4 selected lines of swine over a 4 yr period (2000 through 2003) from PIC (Franklin, KY) were available with 15 pigs per pen. Genetic parameters were estimated by REML with animal models including competition effects. In general, estimates of heritability for direct effects with single-trait or two-trait analyses were in the range of estimates with literature. Estimates of heritability for competition effects were close to zero for all traits. Competition effects, however, would be summed over the number of competitors. Estimates from reduced models indicated that ignoring the small variation due to competition effects could result in changes in estimates of variance components. As an example, excluding competition (genetic and environmental) effects from models appeared to overestimate pen variance. With pens as fixed effects and without environmental competition effects, competition effects are confounded with pen effects. The pattern of estimates of variances, however, can be demonstrated with equivalent models. Furthermore, models including either pen effects or environmental competition effects as random effects are suggested to avoid bias in estimates of genetic variances. Pair-wise analyses between all 5 traits were investigated with bivariate animal models. Convergence was very slow due to much less sparseness in the mixed model equations with competition effects. Estimates of genetic parameters with bivariate analyses were in agreement with single-trait analyses. Antagonisms between direct and competition effects for D110 and DFI were found but based on relatively small estimates of variance due to genetic competition effects. A singular average information matrix appeared in the two-trait analysis including competition effects in the model. In addition, empirical responses to selection for ADG from a weighted selection index were calculated to investigate the relative importance of competition effects and genetic loss due to ignoring competition effects. Results suggested that incorporating competition effects in genetic evaluation might be considered depending on environments that are likely to result in competition among pigs in a pen.