Predicting phenotypic distances using a linear model : the case of varietal distinctness - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Journal of Applied Statistics Année : 2001

Predicting phenotypic distances using a linear model : the case of varietal distinctness

Résumé

Diþerences between plant varieties are based on phenotypic observations, which are both space and time consuming. Moreover, the phenotypic data result from the combined eþects of genotype and environment. On the contrary, molecular data are easier to obtain and give a direct access to the genotype. In order to save experimental trials and to concentrate eþorts on the relevant comparisons between varieties, the relationship between phenotypic and genetic distances is studied. It appears that the classical genetic distances based on molecular data are not appropriate for predicting phenotypic distances. In the linear model framework, we de®ne a new pseudo genetic distance, which is a prediction of the phenotypic one. The distribution of this distance given the pseudo genetic distance is established. Statistical properties of the predicted distance are derived when the parameters of the model are either given or estimated. We ®nally apply these results to distinguishing between 144 maize lines. This case study is very satisfactory because the use of anonymous molecular markers (RFLP) leads to saving 29% of the trials with an acceptable error risk. These results need to be con®rmed on other varieties and species and would certainly be improved by using genes coding for phenotypic traits.
Fichier non déposé

Dates et versions

hal-01222432 , version 1 (29-10-2015)

Identifiants

Citer

Gregory Nuel, Stephane Robin, Claire Baril. Predicting phenotypic distances using a linear model : the case of varietal distinctness. Journal of Applied Statistics, 2001, 28 (5), pp.607-621. ⟨10.1080/02664760120047942⟩. ⟨hal-01222432⟩
77 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More