A common reference population from four European Holstein populations increases reliability of genomic predictions

Genetics Selection Evolution

Research

Alfred G de Vries¹  Adrianus PW de Roos¹  Chris Schrooten¹  Guosheng Su²  Mogens S Lund²  Bernt Guldbrandtsen²  Vincent Ducrocq³  François Guillaume⁴  Sébastien Fritz⁵  Tom Druet⁶  Reinhard Reents⁷  Zenting Liu⁷  Franz Seefried⁷ 

[1] CRV, P.O. Box 454, 6800 AL, Arnhem, the Netherlands;Faculty of Science and Technology, Department of Molecular Biology and Genetics, AU-Foulum, Aarhus University, PO Box 50, DK-8830, Tjele, Denmark;INRA, UMR1313 Génétique Animale et Biologie Intégrative, F-78352, Jouy-en-Josas, France;INRA, UMR1313 Génétique Animale et Biologie Intégrative, F-78352, Jouy-en-Josas, France;Institut de l'Elevage, 149 rue de Bercy, F-75595, Paris, France;UNCEIA, 149 rue de Bercy, F-75595, Paris, France;Unit of Animal Genomics, Faculty of Veterinary Medicine and Centre for Biomedical Integrative Genoproteomics, University of Liège, B-4000, Liège, Belgium;VIT, Heideweg 1, 27283, Verden, Germany;

关键词: Genetic Correlation;    Single Nucleotide Polymorphism;    Protein Yield;    Genomic Prediction;    Somatic Cell Score;   

DOI  :  10.1186/1297-9686-43-43

 received in 2011-05-17, accepted in 2011-12-12,  发布年份 2011

来源: Springer

【 摘 要 】

BackgroundSize of the reference population and reliability of phenotypes are crucial factors influencing the reliability of genomic predictions. It is therefore useful to combine closely related populations. Increased accuracies of genomic predictions depend on the number of individuals added to the reference population, the reliability of their phenotypes, and the relatedness of the populations that are combined.MethodsThis paper assesses the increase in reliability achieved when combining four Holstein reference populations of 4000 bulls each, from European breeding organizations, i.e. UNCEIA (France), VikingGenetics (Denmark, Sweden, Finland), DHV-VIT (Germany) and CRV (The Netherlands, Flanders). Each partner validated its own bulls using their national reference data and the combined data, respectively.ResultsCombining the data significantly increased the reliability of genomic predictions for bulls in all four populations. Reliabilities increased by 10%, compared to reliabilities obtained with national reference populations alone, when they were averaged over countries and the traits evaluated. For different traits and countries, the increase in reliability ranged from 2% to 19%.ConclusionsGenomic selection programs benefit greatly from combining data from several closely related populations into a single large reference population.

【 授权许可】

CC BY   
© Lund et al; licensee BioMed Central Ltd. 2011