Segregation between breeds and local breed proportions in genetic and genomic models for crossbreds

Jón H. Eiríksson*, Guosheng Su, Ismo Strandén, Ole F. Christensen

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

Background: The breeding value of a crossbred individual can be expressed as the sum of the contributions from each of the contributing pure breeds. In theory, the breeding value should account for segregation between breeds, which results from the difference in the mean contribution of loci between breeds, which in turn is caused by differences in allele frequencies between breeds. However, with multiple generations of crossbreeding, how to account for breed segregation in genomic models that split the breeding value of crossbreds based on breed origin of alleles (BOA) is not known. Furthermore, local breed proportions (LBP) have been modelled based on BOA and is a concept related to breed segregation. The objectives of this study were to explore the theoretical background of the effect of LBP and how it relates to breed segregation and to investigate how to incorporate breed segregation (co)variance in genomic BOA models. Results: We showed that LBP effects result from the difference in the mean contribution of loci between breeds in an additive genetic model, i.e. breed segregation effects. We found that the (co)variance structure for BS effects in genomic BOA models does not lead to relationship matrices that are positive semi-definite in all cases. However, by setting one breed as a reference breed, a valid (co)variance structure can be constructed by including LBP effects for all other breeds and assuming them to be correlated. We successfully estimated variance components for a genomic BOA model with LBP effects in a simulated example. Conclusions: Breed segregation effects and LBP effects are two alternative ways to account for the contribution of differences in the mean effects of loci between breeds. When the covariance between LBP effects across breeds is included in the model, a valid (co)variance structure for LBP effects can be constructed by setting one breed as reference breed and fitting an LBP effect for each of the other breeds.

Original languageEnglish
Article number45
JournalGenetics Selection Evolution
Volume55
Issue number1
DOIs
Publication statusPublished - Dec 2023

Bibliographical note

Publisher Copyright:
© 2023, The Author(s).

Fingerprint

Dive into the research topics of 'Segregation between breeds and local breed proportions in genetic and genomic models for crossbreds'. Together they form a unique fingerprint.

Cite this