Genetic progress and inbreeding rate in complex breeding programmes – Applications to sport horses and laying hens

Abstract

The main focus of this thesis is the optimization of breeding programmes. On the one hand a new method to account for the average inbreeding in the design of breeding programmes has been developed. On the other hand two entirely different breeding programmes have been modelled and current optimization approaches are validated using these reference programmes. When calibrating the programmes potential natural and monetary genetic gains as well as the discounted profit are taken into consideration. In the project FUGATO+brain the software ZPLAN for optimization of breeding programmes was re-programmed and further tools were added. The result of the project is the software ZPLAN+. This software enables the user to model and optimise complex breeding programmes. In addition, the software is user friendly and covers all areas of breeding programmes and hence can be used in various contexts. To calculate the average inbreeding and effective population size in complex breeding programmes, a new method has been developed. The method is based on the average kinship in breeding populations. The kinship is defined as the probability that within a group at the same locus two randomly chosen alleles are identical by descent. The calculation of the Kinship is based on the gene flow theory. To validate the method, a sheep population described earlier was used and modified in different ways. Three different scenarios were modelled. The first scenario assumes population growth. In the second scenario it is assumed that the population size is reduced by a bottleneck and then increased again. For the third scenario the population was divided in two parts over a period and then brought together again. The results of this validation exercise show that average inbreeding and effective population size can be calculated in all three scenarios. In a breeding programme for sport horses usage of embryo transfer was validated. A basic breeding programme in ZPLAN+ was modelled, which reflects the current breeding programme of the Hannoveraner Verband e.V. approximately. Using different scenarios, a more rigorous selection on the mares’ side was modelled. The best mares in the breeding programme were used as donor mares for embryo transfer. It was assumed that for selection of the donor mares information of stud book inspection as well as results of a mare performance test are available. The number of mares available for selection, the number of selected mares and the number of born foals per donor mare were varied in order to validate the methodology. The results show that the usage of embryo transfer offers one possibility to increase genetic gain in a breeding programme strongly. However using embryo transfer implies a steep increase in costs for the breeders. The proposed approach for calculating inbreeding showed that the rigorous selection and the intensive usage of donor mares results in an increase of the average inbreeding and consequently a reduction of the effective population size. The third section of this thesis examines the effects of including and using genomic information in a layer breeding programme. A breeding programme to produce 500 million laying hens has been modelled using ZPLAN+ in close cooperation with the Lohmann Tierzucht GmbH. The production of the parents is based on a crossing of four nucleus lines. The conventional selection is based primarily on performance testing of hens of the different lines. In order to utilise the genomic information two different calibration sets were used (500 and 2`000 animals). In a first step, the genomic information of the cocks has been used in addition to all conventional selection criteria. The number of cocks was varied and in a further step it was assumed that the hens are also genotyped. In another scenario, the selection was based on pedigree and genomic information only. It became clear that the generation interval could be strongly reduced in the second variant. The genetic gain could be increased in all modelled variants, but there were differences in individual traits. The implementation of genomic information into layer breeding programmes is connected to a massive increase in costs. Whether the increased genetic gain justifies the increase costs requires a market analysis by the breeding companies.