Abstract
Huiming Liu, Ole Fredslund Christensen, Kevin Byskov, Jón Hjalti Eiríksson, Viktor Milkevych, Jørn Rind Thomasen and Emre Karaman
Recently, crossbred animals have become integral as parents in subsequent generations of dairy and beef cattle systems in Nordic countries, which has raised interest in routine genomic evaluation of these animals. Given that the effects of marker alleles in crossbred animals can vary based on the breed origin of the alleles (BOA), achieving accurate genomic prediction requires a reliable and efficient method for detecting BOA from reference purebred populations, especially in rotational crossbreeding setups. Therefore, this study aimed to evaluate different software and methodologies for their accuracy and efficiency in tracing the BOA using reference purebred populations.
We conducted simulations and evaluated BOA methods with two populations exhibiting different levels of admixture: Nordic Red Cattle (RDC), an admixed breed, and a dairy cross population consisting of Holstein, Jersey, and RDC, similar to those in official Nordic Cattle Genetic Evaluation (NAV).
In Nordic countries, RDC population consists of animals with varying proportions of genetic materials from Danish Red (RDM), Swedish Red (SRB), Finnish Ayrshire (FAY). We identified a total of 304 reference purebred animals by selecting those with a pedigree-based breed proportion of RDM, SRB or FAY greater than 0.9 and conducted PCA analysis on these animals, resulting in three distinct clusters. The number of animals was adjusted to match the real RDC breed proportions in the NAV RDC population: 8 RDM (8.5%), 29 SRB (39.6%) and 38 FAY (51.9%). We phased these animals’ genotypes using Beagle software and used them as the animals in base population and purebred reference populations. We used ADAM simulation software to randomly mate animals for 10 generations and gradually expand the population to approximately 4000 RDC animals by allowing donor scheme with Multiple Ovulation and Embryo Transfer (MOET), matching the real RDC population’s breed proportions and LD structure. Then, we simulated Holstein (HOL), Jersey (JER) and RDC crosses using real HOL and JER haplotypes from DairyCross project, along with simulated RDC haplotypes in the final generation to create crossbred population with the breed proportions similar to NAV crossbred population 0.52 HOL, 0.15 JER and 0.33 RDC. ADAM was updated to enable BOA detections for markers and QTL.
Next, our plan is to trace BOA for these two admixed populations using different software such as AllOr and ChromPainter with varying parameter setups and compare them with real BOA from ADAM simulations. Ongoing analyses will contribute to improvement of genomic evaluation in DairyCross NAV routine genomic evaluation, and the development and implementation of a single-step model combined with BOA.
Recently, crossbred animals have become integral as parents in subsequent generations of dairy and beef cattle systems in Nordic countries, which has raised interest in routine genomic evaluation of these animals. Given that the effects of marker alleles in crossbred animals can vary based on the breed origin of the alleles (BOA), achieving accurate genomic prediction requires a reliable and efficient method for detecting BOA from reference purebred populations, especially in rotational crossbreeding setups. Therefore, this study aimed to evaluate different software and methodologies for their accuracy and efficiency in tracing the BOA using reference purebred populations.
We conducted simulations and evaluated BOA methods with two populations exhibiting different levels of admixture: Nordic Red Cattle (RDC), an admixed breed, and a dairy cross population consisting of Holstein, Jersey, and RDC, similar to those in official Nordic Cattle Genetic Evaluation (NAV).
In Nordic countries, RDC population consists of animals with varying proportions of genetic materials from Danish Red (RDM), Swedish Red (SRB), Finnish Ayrshire (FAY). We identified a total of 304 reference purebred animals by selecting those with a pedigree-based breed proportion of RDM, SRB or FAY greater than 0.9 and conducted PCA analysis on these animals, resulting in three distinct clusters. The number of animals was adjusted to match the real RDC breed proportions in the NAV RDC population: 8 RDM (8.5%), 29 SRB (39.6%) and 38 FAY (51.9%). We phased these animals’ genotypes using Beagle software and used them as the animals in base population and purebred reference populations. We used ADAM simulation software to randomly mate animals for 10 generations and gradually expand the population to approximately 4000 RDC animals by allowing donor scheme with Multiple Ovulation and Embryo Transfer (MOET), matching the real RDC population’s breed proportions and LD structure. Then, we simulated Holstein (HOL), Jersey (JER) and RDC crosses using real HOL and JER haplotypes from DairyCross project, along with simulated RDC haplotypes in the final generation to create crossbred population with the breed proportions similar to NAV crossbred population 0.52 HOL, 0.15 JER and 0.33 RDC. ADAM was updated to enable BOA detections for markers and QTL.
Next, our plan is to trace BOA for these two admixed populations using different software such as AllOr and ChromPainter with varying parameter setups and compare them with real BOA from ADAM simulations. Ongoing analyses will contribute to improvement of genomic evaluation in DairyCross NAV routine genomic evaluation, and the development and implementation of a single-step model combined with BOA.
Originalsprog | Engelsk |
---|---|
Publikationsdato | 22 jul. 2024 |
Status | Udgivet - 22 jul. 2024 |