Genomic selection in mink yield higher accuracies with a Bayesian approach allowing for heterogeneous variance than a GBLUP model

Research output: Contribution to book/anthology/report/proceedingArticle in proceedingsResearchpeer-review

The accuracy of genomic prediction for mink was compared for single-trait and multiple-trait GBLUP models and Bayesian models that allowed for heterogeneous (co)variance structure over the genome. The mink population consisted of 2,103 brown minks genotyped with the method of genotyping by sequencing. Four live grading traits and four traits on dried pelts for size and quality were analysed. GWAS analysis detected significant SNPs for all the traits. The single-trait Bayesian model resulted in higher accuracies for the genomic predictions than the single-trait GBLUP model, especially for the traits measured on dried pelts. We expected the multiple-trait models to be superior to the single trait models since the multiple-trait model can make use of information when traits are correlated. However, we did not find a general improvement in accuracies with the multiple-trait models compared to the single-trait models. Keywords: GWAS, GBLUP, BayesAS, heterogeneous (co)variances
Original languageEnglish
Title of host publicationProceedings of the World Congress on Genetics Applied to Livestock Production, 2018 : Volume Methods and Tools - Prediction 2
Number of pages6
Publication year2018
Article number11.618
Publication statusPublished - 2018
EventThe 11th World Congress on Genetics Applied to Livestock Production - Aotea Centre, Auckland 1010, Auckland, New Zealand
Duration: 11 Feb 201816 Feb 2018
Conference number: 11

Conference

ConferenceThe 11th World Congress on Genetics Applied to Livestock Production
Nummer11
LocationAotea Centre, Auckland 1010
LandNew Zealand
ByAuckland
Periode11/02/201816/02/2018

See relations at Aarhus University Citationformats

Download statistics

No data available

ID: 127392746