Towards a deeper understanding of genetics of dermatitis digitalis in dairy cows through the consideration of housing characteristics, climate and barn emissions in alternative statistical modelling approaches

Abstract

Claw disorders are one of the main reasons for the culling of dairy cows. Due to the lameness that often follows in the herd, the feed intake and, not least, the milk yield of the affected animals decrease. These indirect costs are further increased by the also decreasing fertility of affected animals. In addition, there are the direct costs of veterinary treatment, medication and the need for more care of the lame animals. One of the most frequent claw disorders in this area, occurring worldwide, is dermatitis digitalis, usually better known as Mortellaro's disease. Dermatitis digitalis (DD) is a multifactorial disease in which both the housing environment and the genetics of the animals play a role. In recent years, several studies have been published to clarify the genetic background of claw disorders in general and DD in particular. In this context, genome-wide association studies (GWAS) have been conducted and potential candidate genes associated with claw disorders have been annotated. However, it is still unclear how specific environmental effects influence the estimation of variance components and GWAS. Climatic conditions in particular have changed in recent years and will continue to have an increasing impact on the productivity and welfare of cattle in the future. The Temperature-Humidity Index (THI) is a scientifically proven tool in this context. Heat stress (HS) for cattle begins at a THI > 68. The wide availability of data from weather stations near farms enables appropriate management. In order to identify possible correlations with claw health, SNP x heat stress (HS) interactions for claw disorders were analysed in the first part of this study. This showed that specific SNPs and their annotated candidate genes were only significant under HS or thermoneutral conditions. In the second part of this study, the aspect of interaction effects was taken up again. Here, SNP x housing system interactions were estimated for three DD traits in the housing systems of cubicle housing and compost bedded pack barn. However, very similar genetic parameters were found for the same traits in different environments. As well as negligible genotype x housing system interactions. This indicates that the housing system has only a minor effect on the genetic evaluation of a DD disease. Since cattle farming must adapt to the negative consequences of climate change, on the one hand, and is mentioned as a contributor to climate change, on the other, these aspects should also be reflected in the study design. Therefore, on the one hand, climate influences (wind speed, temperature and humidity) were recorded, and on the other hand, the emissions of ammonia, nitrous oxide, carbon dioxide and methane were measured in the barn. A special ‘climate gas trolley’ was built for this purpose. Using this, it was possible to measure the effects of climate and emissions at the same time in different areas of the barn. In the third part of this study, the effects of the housing environment, cow phenotypes and genomics were placed in an overall context. This approach has been used very rarely in animal breeding to date and not at all in relation to a DD disorder. Structural equation modelling (SEM) was used to address this issue. These models can incorporate both measurable and non-measurable variables into the analyses and have mostly been used in psychology until now. In this study, the models allow for the simultaneous consideration of environmental and genetic effects. The SEM showed that the housing environment has a greater influence on DD than genetic parameters, for example. In summary, the results of this study contribute to a better understanding of the genetic background of claw disorders in general and DD in particular. The results also provide an important contribution to a better understanding and further clarification of genotype-environment interactions. To provide a general introduction to the subject of the dissertation, Chapter 1 will take a closer look at the disease DD. On the one hand, the general course of the disease with its clinical symptoms, therapy and prevention is discussed, but on the other hand, economic aspects and the complexity of possible causes of the disease are also mentioned. The following part of the chapter deals with the housing systems of conventional cubicle housing and compost bedded housing. The chapter ends with various possibilities for modelling and the current knowledge on breeding for disease resistance. Chapter 2 aims at a deeper genomic analysis of the three claw disorders DD, hyperplasia interdigitalis (HYP) and Sole ulcer (SU), as well as their genetic association with other important breeding goal traits. Finally, possible SNP x heat stress interactions for claw disorders are analysed. The study design included 17,264 genotyped, first-lactating Holstein Friesian cows from 50 herds in northeast Germany. The heritabilities of the three claw disorders were estimated using linear and threshold models and were 0.04 and 0.08 for DD, 0.03 and 0.10 for SU and 0.03 and 0.23 for HYP. Genetic correlations estimated in bivariate linear models were consistently positive with the selected conformation traits. This indirectly indicates the need for selection on conformation traits to improve claw health. Genetic correlations with other breeding goal traits showed reduced fertility, poorer udder health and productivity in diseased cows. Genetic correlations were observed between the claw disorders, suggesting a closer genetic relationship. Furthermore, disease-specific candidate genes and genetic associations based on the surrounding SNPs were estimated, which in some cases differed from the genetic correlations. For the SNP x heat stress interactions, significant SNPs were identified on BTA 2,4,5,7,8,9,13,22,25 and 28. The results suggest gene-specific mechanisms for claw disorders only in specific environments. In Chapter 3, a detailed phenotypic characterisation of the claw disorder DD in the two housing environments (conventional cubicle barn and compost bedded pack barn) was carried out to determine possible genotype x housing system interactions. In total, the data set comprised 2,980 observations for the three traits DD-sick, DD-acute and DD-chronic from 1,311 Holstein-Friesian and 399 Simmental cows, 926 of these animals were available for the genomic studies. In total 899 cows were housed in the compost bedded pack barn (1,530 observations) and 811 were housed in the conventional cubicle barn (1,450 observations). The disease prevalence was higher in the cubicle barn than in the compost bedded pack barn. The heritabilities over the entire data set were 0.16 for DD-sick, 0.14 for DD-acute and 0.11 for DD-chronic. A slight increase in heritabilities and genetic variances was observed in the housing environment of the cubicle barn compared to the compost bedded pack barn. Genetic correlations between the same DD traits in the different housing environments were close to 0.80, indicating obvious genotype x housing system interactions. The genetic correlations between the three DD traits ranged from 0.58 to 0.81. The SNP main effects and SNP x housing system interactions were estimated using genome-wide association studies. Some common candidate genes were identified for DD-sick and DD-acute. The genes had direct or indirect effects on disease resistance or immunological processes. The genes ASXL1 and NOL4L were annotated for the SNP x housing system interactions for DD-sick and DD-acute. Chapter 4 analyses the effects of the environmental factors, cow phenotypes and genomics in more detail for the DD traits discussed in Chapter 3. The most relevant housing characteristics were analysed in linear models. The last-squares means for the infection probability were generally lower in the compost bedded pack barn than in the cubicle barn. The genome-wide association studies showed similar Manhattan plots for DD-sick and DD-acute and similar potential candidate genes in each case. Five SNPs were significantly associated with either DD-acute and DD-sick or with DD-chronic and DD-sick. These significant SNPs were then related to phenotypic and genetic estimated breeding values for the DD traits, as well as to production data and housing environmental factors in structural equation models. This showed that the housing environment had a greater influence on the risk of DD infection than genetic parameters, for example. Finally, some important aspects of this work are discussed again in chapter 5. In particular, genomic aspects of claw disorders, the influence of the housing environment and the economic weight of claw disorders in the total breeding value are considered here.