Genetic evaluation of longevity in dairy cattle - A new model for an old trait

Longevity is an important trait in dairy cows, reflecting the overall functionality of a cow. The aim of this thesis was the development of a new model for the routine genetic evaluation of longevity in German Holsteins. To achieve this purpose, different studies were conducted. They are briefly summarized in the following: Chapter 1 provides the reader with background information on base principles of survival theory and on the frame conditions of routine genetic evaluations of longevity in German Holsteins. At the end, the necessity for a new routine evaluation system is defined. With the study in chapter 2, the basis for the new model was developed. It originated from the idea that different periods in the life of a cow belong to different challenges which are related to different functional traits. Binary survival codes were defined for different periods across the first three lactations and modeled as genetically correlated traits. In order to estimate genetic parameters, an excessive estimation of variance components was conducted on data of 1,495,441 cows with two models. With the first model, 18 finely graded periods were considered. Results from this model showed a clear pattern for the genetic background of survival across the first three lactations. Periods with similar genetic background for survival were then merged in the second model, where only nine traits, three for each lactation, were considered. Afterwards, a genetic evaluation was run on data of 7,684,455 cows and estimated breeding values (EBVs) for sires were compared to routine EBVs for other traits. This comparison further justified the approach, showing plausible correlation patterns. In the prototype version of the new routine genetic evaluation of longevity, almost the same periods are used to define different survival traits. These are: survival from calving to 49 d, 50 d to 249 d and from 250 d to the consequent calving. These periods were defined for each of the first three lactations. In chapter 3 it was questioned if models for routine evaluations of longevity should include an effect of age at first calving. The idea for this study arose from the definition of age at first calving (AFC) which is the sum of age at first insemination (AFI), the interval from first to last (successful) insemination (FLI) and gestation length. These traits are all functional traits. In order to investigate if these traits are genetically correlated to survival, variance and covariance components were estimated between AFC, AFI, FLI and survival of different periods of the first lactation as defined in the paragraph above. Data of 721,919 German Holstein cows were analyzed. Estimated genetic correlations of AFC and FLI to survival traits late in lactation were different from zero. As a conclusion, the correction for age at first calving in models for genetic evaluations of longevity should be reconsidered, because it might remove functional genetic variance. Chapter 4 gives a description of the prototype version of the new genetic evaluation system for longevity. The development of this prototype version was part of the project and its results are the basis for chapter 5. The basic idea for chapter 5 was that the differentiated genetic background of survival of different periods should also express in differentiated genome-wide associations. Therefore, this genome-wide association study (GWAS) was performed on deregressed EBVs of 4,849 bulls for the nine survival traits using high-density SNP-marker genotypes. Three different analyzes were performed: (1) a single-marker GWAS (2) a gene-based GWAS and (3) a gene-based mixed model, where gene regions with significant associations identified from (2) were modeled as random. Eight regions on chromosomes 5, 6, 7, 14 and 18 showed significant associations to at least one of the survival traits. Different patterns were observed for the strengths of association among the survival traits. These were in most cases plausible when compared to results from other studies. The study in chapter 5 justifies the results of chapter 2 from a genomic point of view and lays the foundation for further research on this topic. Results from this study may also be valuable when designing models for genomic evaluations of longevity in dairy cows. In chapter 6, important topics that were not covered by chapters 2 to 5 were further highlighted and discussed in detail. It gives insights into different methods for the construction of an index EBV from nine survival traits. Potential for further research from observations during the study in chapter 5 is illustrated. At the end, a short prospect of the future is given for the longevity trait, the overall indicator of functionality