Modeling and Optimization of Animal Breeding Programs
Cumulative thesis
Date of Examination:2024-03-20
Date of issue:2024-11-22
Advisor:Prof. Dr. Henner Simianer
Referee:Prof. Dr. Jens Tetens
Referee:Prof. Dr. Steffen Weigend
Files in this item
Name:Dissertation_Lisa_Büttgen.pdf
Size:4.17Mb
Format:PDF
Abstract
English
Simulation software can be used to model and optimize breeding programs. In the current work, the “Modular Breeding Program Simulator” (MoBPS) was applied to demonstrate this on three examples from layer, horse and pig breeding. For this purpose, the R package MoBPS as well as the corresponding steadily evolved user interface MoBPSweb were used. Simultaneously, the user interface can also be used to describe breeding programs in a unified way as cohorts of individuals and breeding actions. The main focus of the simulations in this thesis is to compare different scenarios for the selection against exemplary diseases and welfare-relevant behavioral traits as well as to mimic breeding programs with a fairly realistic age structure including overlapping generations. To address this, different scenarios of a layer breeding program were simulated. Besides the analysis of different scenarios for the selection, including scenarios with genomic selection and optimum contribution selection, the focus was on shortening the generation interval through the use of two time-shifted subpopulations with shortened generation intervals for males and unchanged generation intervals for females. It turned out that shortening the generation interval only on the male side led to an increase in genetic gains, but also to a substantial increase in inbreeding rates. Furthermore, the simulations showed that genotyping of hens in combination with genomic selection led in comparison to pedigree-based BLUP to increases in genetic gains as well as much lower inbreeding rates. In terms of selection for health traits, the possibilities of selection against partially genetically caused diseases are also being evaluated on the example of osteochondrosis dissecans (OCD) in horse breeding. For this purpose, the breeding program of the Hanoverian riding horse population was mimicked including a realistic age structure. In different scenarios threshold selection, index selection or genomic index selection were simulated as potential approaches to breed against OCD. The simulations showed that both a rigorous threshold selection against the occurrence of OCD and the integration of the trait into a selection index at the stage of stallion licensing in combination with the frequency of breeding use based on a selection index including breeding values for OCD traits are promising scenarios. For the reduction of OCD, the sole integration of OCD in a selection index was less effective than both previously mentioned scenarios. In scenarios with a higher reduction in the disease trait, there was a slightly reduced increase in performance. The use of auxiliary traits is particularly useful for traits that are difficult or costly to record, such as tail biting. To address the reduction of tail biting in pigs by indirect selection of an auxiliary behavior trait, different possible correlations of such a trait with tail biting, as well as different index weights of the auxiliary trait, were simulated. In the simulations, tail biting could be sufficiently reduced by the integration of an at least moderately correlated auxiliary trait, even with a relatively low weight on the auxiliary trait. There was no substantial slowing of breeding progress in the performance traits even when a slightly higher weight was put on the auxiliary trait to accelerate the reduction in tail biting. Overall, stochastic simulations of breeding programs turned out to be a useful tool for weighing up different options for optimizing breeding programs. A variety of scenarios can be compared, including those that serve to delineate effects but are not considered for implementation in reality, as well as to investigate of the effects of possible values for e.g. correlations that are not yet sufficiently known.
Keywords: animal breeding; breeding programs; simulation; MoBPS