Development of a genetic tool for determining breed purity of cattle

Breed registries have been established for livestock species to maintain the purity of breeds and to document the ancestry of animals. However, a significant number of animals are unregistered with no or incomplete pedigree data and uncertain ancestral breed origin. Although many local livestock breeds are "at risk" on the basis of the number of purebred breeding females in a breed registry, there is often also a reservoir of unregistered animals that may belong to the same breed. However, due to the missing pedigree it is not possible for breed societies or herd books to include those animals in their breeding program for purebred animals. A genetic test was developed to unequivocally determine the breed origin of cattle without pedigree data. Such a test will open up the possibility to incorporate animals without pedigree data in the breed registry that turn out to be purebred based on the test results. In this study we developed and validated such a test. Genotype data (50k SNP array) were used to compose reference populations for six local Dutch cattle breeds. The combination Principal Component Analysis and Random Forest was used to perform SNP selection. A total of 133 informative SNPs were selected to determine breed composition of individual animals. Overall, 82.0% of the animals in the test population are correctly assigned to the breed in question. For Dutch Red and White Friesian and Deep Red Cattle we suggest that if an animal has a percentage for its own breed < 0.775 to use the combined percentage of two breeds (Deep Red Cattle with Meuse-Rhine-Yssel and Dutch Red and White Friesian with Dutch Friesian). Using this criteria 88.9% (104 out of 117) of the animals in the test population is correctly assigned. The developed test was successful and will be implemented in practice to identify (partly) unregistered individuals as being purebred (or not) for one of the Dutch local cattle breeds.