Abstract

Four different mixed models were used to estimate additive and dominance genetic values of animals from simulated data. Five generations were produced allowing inbreeding using a genetic model with a large number of loci (64) and two different gene frequencies (0.5 and 0.8). There was complete dominance at all loci. Population size at each generation was 40 and 5 males were either randomly selected or selected on phenotype and mated to 20 females. In generation 5, average inbreeding was 0.08. A model with an additive relationship matrix (A), dominance relationship matrix (D) and inbreeding coefficient (F) as a covariate gave unbiased prediction of additive and dominance values and unbiased estimation of inbreeding depression under random as well as phenotypic selection. A model which included A but ignored D and F underestimated predictors of additive genetic values under both random and phenotypic selection and a model that included A and F but ignored D overestimated additive genetic values with phenotypic selection. A model which ignored A and D and included only F seriously underestimated inbreeding depression when phenotypic selection was practised.
 

P. Uimari, B. W Kennedy

Proceedings of the World Congress on Genetics Applied to Livestock Production, Volume XIII. Plenary lectures, molecular genetics and mapping, selection, prediction and estimation., , 297–300, 1990