Innovative combination of all sources of information for production traits in Slovenian Brown Swiss


  • Klemen Potočnik University of Ljubljana, Biotechnical Faculty
  • Jeremie Vandenplas University of Liege National Fund for Scientific Research
  • Marija Špehar Croatian Agricultural Agency
  • Nicolas Gengler University of Liege
  • Gregor Gorjanc University of Ljubljana, Biotechnical Faculty The Roslin Institute and Royal (Dick) School of Veterinary Studies, The University of Edinburgh


single-step genomic BLUP, combination, external, genomically enhanced EBV


Slovenian Brown Swiss is a small population with genetic improvement based on its own breeding program supplemented with imports from other populations. Routine national genetic evaluations for milk, fat, and protein yields are computed from all available national phenotypic and pedigree data. At an international level, a Multiple Across Country Evaluation (MACE) is performed by Interbull to aggregate estimated breeding values (EBV) for international sires across different populations into a single Slovenian ranking. Additionally, a genomic evaluation for many sires is now routinely computed at an international level through the InterGenomics (IG) project. Phenotypes used for this genomic evaluation are deregressed MACE EBV which generate genomically enhanced EBV (GEBV) for all genotyped sires. However, national evaluations are not influenced by these international evaluations and, therefore, may be less accurate and even biased because foreign data used to select foreign sires are not used at the national level. Therefore, an integration of international evaluations back into the national evaluations is required to use the available information in an optimal way for both bulls and cows. The aim of this study was to assess the potential of an innovative Bayesian approach, based on a single-step genomic Best Linear Unbiased Prediction, that combines national data for milk, fat and protein yields with the IG genotypes and information (i.e., GEBV and reliabilities). Because IG information considers genotypes and also MACE information, which also includes national information, double-counting of contributions due to records and due to relationships had to be considered. The integration of IG genotypes and information showed an increase of reliability for the three traits, especially for all IG sires. For example, for IG sires with progeny with national records, the integration led to an average increase of reliability of > 0.10 points for milk yield, in comparison to their average national reliability. For the IG sires without progeny with national records, an average increase of reliability of >0.74 points was observed for the same trait. An average increase of reliability of > 0.05 points was also observed for animals with a reliability <0.30 and sired by genotyped IG sires and with progeny with records. Finally, this approach has the potential to simultaneously combine national data and IG genotypes and information. Furthermore, while it was not implemented in this study, this approach has the advantage to allow the consideration of genotypes of other non-IG animals (e.g., cows).

Author Biography

Klemen Potočnik, University of Ljubljana, Biotechnical Faculty

Department of Animal Scince

Computing Centre leader