Breeding and Genetics: Genetics Methodology 1

Simulation of the model of Muir and Schinckel containing genetic competition effects was initiated to determine how well REML could untangle variances due to direct and competition genetic effects and pen effects. A two-generation data set was generated with 6 unrelated males, each mated to 5 unrelated females to produce 300 progeny from which 30 females (one per mating in previous generation) were mated to 6 unrelated males to produce 300 more progeny. Progeny were randomly assigned 6 per pen to 50 pens per generation. Parameters were Vg, Vc, Cgc, Vp and Ve representing direct and competition genetic variance with covariance, and pen and residual variance. Eight statistical models were used to analyze each of 400 replicates of 16 sets of parameters. Both Vg and Ve were fixed at 16. Values of Cgc were -2, -1, 0.1, 1 and 2. Values of Vc were 1 and 4 and of Vp were 0.1, 1, and 10. With the full model, average estimates resembled true parameters except that Vp was consistently overestimated when small (0.1 and 1) which, in turn, slightly changed other estimates. The most unexpected result was overestimation of Vp when Vc and Cgc were ignored in the analysis. The overestimation depended on Vc and number of competitors in common between records in a pen. The upward bias was greater when Cgc was positive than when negative. For example, with Cgc = 2, Vc = 4 and Vp = 0.1, mean estimate of Vp was 20.4 when Cgc and Vc were dropped from the model and 15.3 when Cgc = -2. When Vp was ignored, estimates of both Cgc and Vc increased proportional to Vp. Also Vg increased more with greater Vp. Another unexpected result was that when pen was considered fixed for the analysis, sampling standard errors of estimates of Cgc and Vc were reduced generally by factors of 2 to 30. These results suggest that a high estimate of pen variance may indicate genetic competition effects are important and that ignoring pen effects will bias estimates of other components.