Premium
On a Mixed Linear Model when the Data are Subject to Selection
Author(s) -
Im S.
Publication year - 1990
Publication title -
biometrical journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.108
H-Index - 63
eISSN - 1521-4036
pISSN - 0323-3847
DOI - 10.1002/bimj.4710320320
Subject(s) - selection (genetic algorithm) , generalized linear mixed model , best linear unbiased prediction , mixed model , trait , linear model , statistics , random effects model , econometrics , model selection , variance (accounting) , mathematics , computer science , artificial intelligence , medicine , meta analysis , accounting , business , programming language
Most data available to animal breeders are subject to selection. We consider estimation and prediction in a mixed linear model for two traits when selection acts on the first trait and data for the second trait are available on selected individuals only. Difficulties of applying best linear unbiased predictors accounting for selection (HENDERSON, 1975) are highlighted. Using inferences based on the likelihood, it is shown that selection can be ignored in estimating variance components and fixed effects and in predicting random effects.