Multiple and Single Trait Analyses for Estimating Genetic Parameters in Simulated Populations Under Selection

Estimation of Genetic Parameters for Milk and Fat Yields of Dairy Cattle in Spain and the United States

“…In this study, milk and fat were highly correlated in the Spanish data, although selection intensity was not expected to be high. In contrast to Walter and Mao's simulation (24) for which a decrease in estimates of residual variance components was reported for single-trait as compared with multiple-trait analyses, estimates of residual components for the Spanish data did not decrease if analyzing milk and fat yields independently rather than jointly but instead increased slightly. Estimates of phenotypic variances were similar for both withincountry analyses and for joint analyses.…”

“…The decrease in the sire variance component for the Spanish data set did not have a clear explanation. Simulation studies by Walter and Mao (24) and Pollak et al (18) and Meyer and Thompson's (15) study with field data found that estimates of genetic variances tended to decrease if correlated traits with selection on one trait were analyzed as single traits rather than as a multiple trait. Magnitude of decrease was related to selection intensity and correlation between traits.…”

Estimation of Genetic Parameters for Production and Reproduction in Finnish Ayrshire Cattle

“…To avoid possible bias due to the selected data (30,32), observations with sequential trait combinations starting with NRR-H were utilized. When an animal was missing data for any of the traits, the missing observations were replaced with a dummy value of 0 and nulls were substituted for corresponding rows of X and Z.…”

Correcting within-family pre-selection in genetic evaluation of growth—A simulation study on rainbow trout

“…Similarly, the accuracy of EBVs was considerably lower for the selected data compared with random sampling (Appel et al, 1995(Appel et al, , 1998. In contrast to within-family selection, a simulation study on cattle breeding in which threshold selection across a population was practiced on an early trait showed reduced genetic and residual variance estimates for a later trait, when selection was not accounted for (Walter and Mao, 1985). Such selection makes the half-sib families of sires more similar, reducing genetic variance in a sire model analysis (Pollak et al, 1984).…”

“…the records of the selected trait from all individuals are included in the analysis (Henderson, 1975). Simulations studying the effects of pre-selection in beef and dairy cattle have demonstrated that the (restricted) maximum likelihood estimates are almost free of bias when multivariate analysis with enough information on the culled animals is used (Ouweltjes et al, 1988;Pollak et al, 1984;Rothschild et al, 1979;Walter and Mao, 1985).…”