Selection on selected records 英文参考文献.docVIP

I.N.R.A., Laboratoire de Biometrie, Centre de Recherches de Toulouse, chemin de Borde-Rouge, F 31320 Castanet- Tolosan . The problem of selecting individuals according to their additive genetic values and of estimating those values, is considered. It is assumed that the selection is based on a vector of observations made on a group of individuals which were themselves selected according to 1 certain vector of observations. An optimal selection rule applicable irrespective of the distribution of the random variable nvolved in the setting is derived. In particular, it is shown that the restrictions regarding the ,H out by ENDERSON can be relaxed. Predictor) pointed Key-words : Selection, mixed models, BLUP. On considère le problème de la sélection d’individus pour leurs valeurs génétiques additives et de l’estimation de ces valeurs. La sélection est basée sur un vecteur d’observations faites ensemble d’individus eux-mêmes issus d’une sélection sur un certain vecteur d’obser- vations. soit la distribution des variables que les contraintes d’utilisation du BLUP (meilleur prédicteur linéaire sans biais) proposées par Henderson, peuvent être atténuées. aléatoires de l’expérience. En particulier, on montre Mots-clés : Sélection, modèle mixte, BLUP. Animal and plant breeders are often faced with the problem of choosing items, e.g. sires or varieties, among set of available candidates. Generally, selection is a H ENDERSON (1973, 1975), in a multi- based on a vector of observations made on these or other items which were themselves selected according to another vector of observations. Therefore, it is important to develop a selection rule that is optimal in some sense.  variate normal setting, showed that if certain conditions related to fixed parameters in a linear model describing the observations are met, then the best linear unbiased predictor (BLUP) eliminates the bias resulting from the previous