In poultry breeding programs, the final product is usually a crossbred animal, and selection takes place in their purebred parental lines. Breeders therefore need to know the correlation between purebred and crossbred performance (r_pc), so they can predict the response to selection in crossbreds. In addition, knowing r_pc can help breeders to make decisions about whether data should be collected on crossbred animals to improve breeding values. But how should r_pc be estimated?
The genetic correlation between purebred and crossbred performance (rpc) is typically estimated with statistical models that use information on relationships between purebred and crossbred animals. Such relationships can be estimated either from pedigree or genotype information, and these information sources may lead to different estimates of rpc. In addition, with genotypes, estimates of rpc may be different when the breed-of-origin of alleles is either considered or ignored.
Comparing estimates of rpc
Researchers from Wageningen University & Research have investigated differences in estimated rpc from pedigree-based and genotype-based models, and they have tried to identify which approach gives the most accurate estimates. Their results showed that genotype-based models resulted in larger estimates of rpc than pedigree-based models, whereas considering the breed-of-origin of alleles did not substantially affect estimates of rpc. The researchers concluded that genotype-based models should be preferred over pedigree-based models, because those estimates had smaller standard errors. Considering the breed-of-origin of alleles may improve estimates of rpc in situations where the parental lines are more closely related than they were in the present study.
Implications and future plans
The aforementioned study used data on body weight in broiler chicken as an example. The results showed that body weight traits in purebreds and crossbreds are genetically very similar, with estimates of rpc ranging from 0.80 to 0.96. It is therefore not expected that the accuracy of breeding values for crossbred performance will improve when data is collected on crossbred instead of purebred animals. A detailed investigation of the benefit of using crossbred data for genomic prediction will be investigated in a follow-up study.
This research was supported by the Netherlands Organisation of Scientific Research (NWO) and the Breed4Food consortium partners Cobb Europe, CRV, Hendrix Genetics, and Topigs Norsvin. Data was provided by Cobb Europe.