Comparison of models for genetic evaluation of scrotal circumference in crossbred bulls.

The aim of this study was to evaluate the effect of including concomitant body weight and(or) a random dam effect in genetic evaluation models on variance component estimates and standard error of prediction for scrotal circumference (SC) at 6, 8, 10, and 12 mo. Variance components and average standard errors of prediction were compared under models differing in either the number of related traits (M11 [SC], M12 [SC and BW]) or an uncorrelated random dam effect (M21 [SC], M22 [SC and BW]) using records on 1,547 bull calves. In a single-trait model (M11), estimates of direct heritabilities (h2a) for SC were .45, .49, .57, and .66 at 6, 8, 10, and 12 mo, respectively. In a two-trait model (M12), h2a were similar to those in M11 model. In M21, h2a for SC were .37, .42, .54, and .65, whereas the proportions of phenotypic variance due to dams (d2) were .12, .11, .04, and .02 at 6, 8, 10, and 12 mo, respectively. Similarly, in M22, h2a for SC were .36, .44, .56, and .65 and d2 were .13, .10, .02, and .02. Standard errors of prediction for SC EBV from M22 were reduced by 2.86, 1.21, 3.02, and 1.99% relative to M21 and by 6.45, 2.70, 2.72, and 1.21% relative to M11 at 6, 8, 10, and 12 mo, respectively. Standard errors of prediction for SC EBV from M12 were reduced by .06, .73, 1.56, and .87% relative to M11 at 6, 8, 10, and 12 mo, respectively. The importance of the dam effect decreased with age for both SC and BW. These results demonstrate that a two-trait (SC and BW) animal model would result in more accurate evaluations of yearling SC EBV in beef cattle than a single-trait model.