Aarhus University Seal / Aarhus Universitets segl

Biology informed genomic selection for purebred and crossbred cattle

Projekter: ProjektForskning

Se relationer på Aarhus Universitet


Breeding values can be estimated using high-throughput genotyping technologies and dense marker maps through genomic selection (GS) (Meuwissen et al. 2001). GS relies on the existence of linkage disequilibrium (LD) between genetic markers and quantitative trait loci (QTL). The effect of each allele is estimated by fitting a model including both phenotypic and genotypic records. It has been successfully implemented in single-breed evaluations (VanRaden & Sullivan 2010; Lund et al. 2011) but the results when using crossbred information vary depending on the breeds and the traits included in the models (Zhou et al. 2014). GS can be challenged in admixed and crossbred populations since the allele substitution effects may differ between breeds. It means that the same allele of a given marker might have a different effect on the crossbreed phenotype according to its breed origin. These differences in effects may be attributable to several causes. First, different levels of LD between SNP alleles and a QTL in the populations. Second, the backgrounds might differ, meaning that the functional variation underlying a QTL is not segregating for the different breeds or epigenetic interaction due to other genes modifying the QTL effect. Third, environmental differences in raising of the animals from each breed may lead to different QTL effects, i.e., genotype by environment (G×E) interactions might be present (Bastiaansen et al. 2014; Vandenplast et al. 2016). The presence of G×E interactions is marked when animals are kept in different countries or climates.
Several simulation studies have included breed origin of allele in the models (Ibanez-Escriche et al. 2009; Esfandyari et al. 2015) assuming that the origin of allele is known. However, estimating the SNP allele effect using real data requires the purebred origin of allele in crossbreds to be determined. Software that determines the breed origin of genome segments accurately is required. Considering the parental origin of alleles in prediction models is expected to improve predictions for and admixed breed as the Nordic Red (which is an admixture of Finnish Ayrshire, Swedish Red, Danish Red, American Brow Swiss, and Holstein) and for a large crossbred beef × dairy Irish cattle population.
With the increase in the exchange of genetic material, usually many daughters of the top bulls are raised in very different environments, such as different countries. G×E may affect their performance. For the case of crosses between Danish and Bos indicus breeds in Indian conditions, it is expected to observe large G×E. Multi-traits models to account for G×E would improve the prediction of individual’s performance in Indian conditions.
Furthermore, the use of epigenetic and genetic expression data could be incorporated into G×E models to improve predictions through knowledge of differentially expressed genes or regions in Denmark and India.
The main aim of the project is to improve predictions models for multibreed populations in different scenarios of breeds, environments and population size. The proposed approach takes advantage of (i) use large Nordic, Irish and Indian data sets of genotyped bulls and cows of multiple breeds, (ii) use sequence information from the 1000 Bulls Genome Project (Daetwyler et al. 2014), (iii) using the improved annotation of the cattle genome in a new EU project: BovREG. As a general result, it is expected to have improved genomic selection for admixed breeds and crossbred performance.
Effektiv start/slut dato01/10/201930/09/2022

ID: 218204020