New models evaluate animals correctly even with incomplete data

New genetic evaluation models known as ‘single-step genomic evaluation models’ predict genetic merits of animals without bias even with incomplete data. This is the main conclusion from an investigation carried out by researchers working at the Animal Breeding and Genomic Group of Wageningen university and Research. 


Selection of animals takes place in multiple stages

In animal breeding, selection of animals to produce the next generation is just like natural selection - it is the survival of the fittest! Only that this time around the fittest animals are determined by animal breeders instead of nature. Depending on the purpose of keeping the animals (such as milk or beef production in cows and egg or meat production in chickens), different animals are selected. Genetic merit of an animal at the time of selection (compared to other animals) is what determines whether the animal would be selected or not. Animal breeders rely on data from large numbers of animals to make these selection decisions. Sometimes, the data available are not as complete as animal breeders would want them to be. An example of this situation is caused by Preselection. In several cases, the ideal moment of selection is when some performance records are available on animals. Raising all animals to this point in time, and recording their performance, is costly. Therefore, an initial selection step is often made early in life of the animals, using either parental or genomic information. This is called preselection, and it is known to cause bias to genetic merits of preselected animals when they are subsequently evaluated to decide whether they will participate in producing the next generation or not.

Subsequent evaluation with old versus new models

Until recently, subsequent evaluation of preselected animals was done with genetic evaluation models that only use pedigree and phenotypic information. These models are not able to account for the fact that data available after preselection are incomplete, and so they usually underestimate the genetic merits of these animals. So-called single step genomic evaluation models combine pedigree, genomic and phenotypic information in one go to evaluate genetic merits of animals. Our results show that these models estimate genetic merits of animals without bias, even after preselection.

The research was published in the journal ‘Genetics Selection Evolution’. Read the full article for more information. 

Contact person: This email address is being protected from spambots. You need JavaScript enabled to view it.