4.1. Genetic evaluation with BLUP
The use of the BLUP Animal Model is referred to as “Genetic evaluation” and its outcome, the “breeding values”, refers to the probability that the progeny of the selected individuals will be above or below the population average for a certain considered trait.
Genetic evaluation aims at assigning a genetic value to each animal with the goal of ranking animals and selecting animals with the best genetic values. Compared to other livestock which undergo genetic improvement, honey bees have peculiar genetic and reproductive characteristics (haplo-diploid sex determination, arrhenotoky, polyandry) which make simple appliance of the BLUP Animal Model not appropriate (difficulties in calculating the numerator relation matrix, which links information from related colonies (Bienefeld et al., 1989; Fu-Hua and Sandy, 2000)). However, the main methodological problem is that the colony’s performance and behaviour result from the interaction between the queen and worker bees. Thus, a trait measured in the honey bee colony is the result of the combined activities of the queen (maternal effect) and workers (direct effect). Bienefeld and Pirchner (1990) found queen and worker effects to be negatively correlated, which strongly hinders selection response (Willham, 1963). Therefore, the BLUP animal model approach was modified to consider worker and queen effects and the negative correlation between them (Bienefeld et al., 2007).
Genetic evaluation via BLUP combines the phenotypic data of the animal itself with data of related animals to rank them according to their (environmentally adjusted) genetic merit. Therefore, this approach needs the individual results of performance tests of all animals and the genetic relationship (pedigree information) between them. All this information must be combined in an appropriate database.
The requirements for the database are the following:
- Controlled (i.e., password-protected) access for data input.
- Software-assisted checking for coherence with existing information, outliers, and logical inconsistencies.
- Clear definition of access rights if several people have written access (e.g. breeder and administrator of a breeding association).
- Data format should fit the requirements of the of the genetic evaluation software.
Open access for all users regarding
the results of the genetic evaluation.
At the moment, just one international database for the honey bee fulfils these requirements (www.beebreed.eu), and so its specifications have been chosen as a standard.
Most breeders use the database not only for efficiently making data of their colonies available for genetic evaluation, but also for running their private studbook. Not all entries of the studbook (e.g. day of birth, tag colour of queen, etc.) are needed for genetic evaluation. To adjust for the environmental effect, information concerning the contemporary group is of central importance. A contemporary group comprises all colonies tested at the same location and management conditions at the same point in time. For genetic evaluation, the contemporary group is formed by combining the following variables: year of birth, login ID of the tester (who is not necessarily the breeder), and a code for the apiary belonging to the tester (one tester may run several apiaries). Ten to 15 colonies per apiary are needed to be able to correctly adjust for the environmental effect of an apiary. However, fewer colonies per apiary are accepted for genetic correlation, but then the colony information at these apiaries is downgraded. Genetic evaluation requires genetic links within the population and is promoted by the simultaneous testing of the different genetic origins (of the same race) at each apiary.
For the reasons explained above (reproductive peculiarities of honey bees) and in contrast to other species, the full pedigree specification in the database used for genetic evaluation consists of the identification number of the (actual) queen, of her mother, of her mating partner, and NOT her father. This model is adapted to the breeding scheme according to which a single drone line is used: a mother queen is selected from whom a group of queen daughters is reared, which will be used for drone production (Ruttner, 1988). The paternal descendent of each queen needed for genetic evaluation is (software-assisted) generated by using pedigree information of her mother. For each drone producing sister group, a dummy father is inserted into the pedigree. The identification number of the mother is a mandatory field in the database, but not for the mating partner, because controlled single-line mating is not adopted by all associations. Pedigree data is combined with performance data for genetic evaluation.