Browsing by Person "Schneider, Helen Hiam"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Publication
    Detailed genomic analysis of correlation and causality between milk production and health traits in German Holstein cattle using high-dimensional genomic data and novel statistical methods
    (2024) Schneider, Helen Hiam; Bennewitz, Jörn
    Adverse side effects of high milk production on animal health have been mentioned frequently. They are compromising animal welfare, the farmers` economy, and the ecological footprint as well as the social acceptance of milk production. Consequently, many countries started to include functional traits into their breeding goal a few decades ago. The intention is thereby to avoid putative undesirable side effects of high production and to improve the cows` health in the long term through genetic gain. Indeed, positive genetic trends for various functional and health traits have been described in the literature. At the same time, the genetic trend of milk production traits remained positive. In general, sustainable genetic gain requires an appropriate weight of the individual traits in the selection index and a comprehensive understanding of the traits` genetic architecture and their interrelationship. This can be facilitated by recent innovations that enable the widespread availability of whole genome sequence (WGS) data. WGS data contains genomic information about millions of SNPs, derived either from sequencing or from imputing lower density SNP chip data to sequence level. Using external information about these sequence variants in genomic analyses, e.g., concerning their function during transcription and translation, has been shown to reveal additional knowledge about biological and molecular mechanisms shaping complex traits. Hence, applying external information to estimate genetic correlations might help to dissect the traits` interrelationship in more detail. Additionally, going beyond global genetic correlations, this is, reflecting the shared genetic effect throughout the genome, to the local scale, this is, the genetic sharing in specific genomic regions, would be an alternative to provide novel knowledge about the extent and direction of the shared genetic effect and its localization in the genome. The expected information is desired to understand and to avoid potential detrimental effects of selection decisions on animal health. Moreover, moving away from correlation towards causation would enable to predict the impact of management decisions and external interventions. The aim of this thesis was to scrutinize the genetic connection between health and milk production traits in dairy cattle using a large sample of 34,497 German Holstein cows with pedigree, 50K SNP chip, and imputed WGS data consisting of ~17 million variants. To this end, standard quantitative genetic analyses were augmented by a set of novel approaches to detect genomic regions with a substantial genetic effect on several traits and to investigate causal associations. Chapter one applied the 50K chip data to estimate additive genetic and dominance variance components for the milk production and health traits. This was done since substantial nonadditive genetic effects for functional traits have been mentioned in the literature, whereas little is known about these effects for the health traits examined in this thesis. It was demonstrated that the contribution of the dominance variance to the phenotypic variance was rather small for all traits. However, regarding the health traits, the contribution of the dominance variance to the genetic variance was almost as high as, and sometimes even higher than the contribution of the additive genetic variance. In addition, significant inbreeding depression was found for the milk production traits. Chapter two consisted of three steps. First, pedigree-based heritabilities of and global genetic correlations between milk production and health traits were estimated. Most heritabilities of the health traits and their genetic correlations with the milk production traits were low, whereby the genetic correlations were in an unfavorable direction. Next, genome-wide association studies (GWAS) were performed for each trait utilizing the 50K chip data to generate summary statistics. The summary statistics are required as input data for the last step that applied a tool to detect shared genomic regions. Genomic regions simultaneously affecting milk production and health traits were identified for each trait combination, of which some also had a sign in the favorable direction. This chapter confirmed the advantage of scrutinizing global genetic correlations down to the local scale. Chapter three utilized the 50K chip as well as the imputed WGS data. The latter was thereby divided into 27 subsets depending on the variants` functional and evolutionary annotation, e.g., as gene expression quantitative trait loci or selection signature. Heritabilities of and genetic correlations between milk yield and several health traits were estimated for the 50K chip and each of the 27 subsets. The results indicate that the 50K chip appears to be sufficient to explain the genetic variance of the investigated traits, whereas it seems to underestimate their genetic covariance. Furthermore, the importance of alternative splicing for the (co-)variation of quantitative traits and the important role of the negative energy balance causing the unfavorable side effects of high production on animal health has been confirmed. Chapter four was a Mendelian randomization (MR) analysis. Here, the causal effect of milk yield on a set of health traits was examined using a method that is based on summary statistics. In this chapter, the summary statistics were generated using the imputed WGS data. Unfavorable causal effects of milk yield on most health traits were identified that were strongest for mastitis and digital phlegmon. This indicates potential detrimental consequences for these traits with increasing milk yields, owed to selection decisions or inappropriately chosen weights in the selection index. The general discussion is addressing the negative side effects of high production on animal health with special focus on the negative energy balance. Moreover, including feed efficiency and resilience indicator traits into the breeding goal is discussed with respect to the results reported in the previous chapters. Besides, additional information about the methodology of MR analyses and the results of a MR analysis investigating the causal effect of protein and fat yield on the health traits are presented and debated. The general discussion ends with practical implications of the results regarding hologenomic selection strategies and strategies including functional information in genomic prediction.