Abstract :
Data from 2032 Uruguayan Aberdeen Angus cows under extensive management and recording practices were analysed with
Bayesian threshold-liability sire models, to assess genetic variability in calving success (CS), defined as a different binary trait
for each of the second (CS2), third (CS3) and fourth (CS4) calving opportunities. Sire (herd) variances ranged from 0.08 to 0.11
(0.10 to 0.20) and heritability from 0.27 to 0.35, with large credibility intervals. Correlations between herd effects on CS at
different calving opportunities were positive. Genetic correlation between CS2 and CS4 was positive (0.68), whereas those
involving adjacent calving opportunities (CS2–CS3 and CS3–CS4) were negative, at 20.39 and 20.54, respectively. The residual
correlation CS2–CS3 was negative (20.32). The extent of uncertainty associated with the posterior estimates of the parameters
was further evaluated through simulation, assuming different true values (20.4, 20.2, 10.2 and 10.4) for the genetic
correlations and changes in the degree of belief parameters of the inverse Wishart priors for the sire covariance matrix.
Although inferences were not sharp enough, CS appears to be moderately heritable. The quality of data recording should be
improved, in order to effect genetic improvement in female fertility.
Keywords :
calving success , genetic parameters , Threshold models , Bayesian theory , beef cows