Browsing by Subject "Tierzucht"

Now showing 1 - 12 of 12

Analysis of phosphorus utilization using the host genome and microbiota variability in Japanese quail
(2021) Vollmar, Solveig Deniece; Bennewitz, Jörn
Phosphorus (P) is an essential element for growth and performance of avian species. It is predominantly bound as phytic acids and salts (phytate) in plant seeds. Phytases and other phosphatases can harness P by cleaving P groups. Nonruminants have low endogenous phytase activity in the gastrointestinal tract, and thus, the requirement of this element is not met from exclusive plant-based diets. Therefore, mineral P or phytase enzymes are supplemented in poultry feed. Due to the finite quantities of high quality mineral P worldwide, it is of great economic interest. P supplementation is increasingly causing environmental problems. Past studies investigated the P utilization (PU) of different poultry species. They revealed a high phenotypic variation in PU among individuals. Moderate heritabilities indicates that breeding for this trait is in principle possible. The overall aim of this thesis was to gain a deeper understanding of the variability of P utilization in relation to host genetics, ileal microbiota composition and their interaction in the model species Japanese quail. The objective of chapter two was to verify whether variation in PU in quail is a heritable trait conditioned by a few quantitative trait loci (QTL) with detectable effects. For this purpose, individuals were genome-wide genotyped with a 4k SNP chip, and a linkage map was generated. Based on this map, QTL linkage analysis was performed using multimarker regression analysis in a line-crossing model to map QTL for PU. We identified a few QTL regions with significant effects. Among them was a QTL peak at Coturnix japonica chromosome (CJA) 3 for PU. Several genes were found in the region surrounding this peak, which requires further functional gene analysis. Based on these results, we hypothesized that these traits are polygenically determined due to several small QTL effects, which we could not detect significantly. The overlap of the QTL regions indicated linkage of the traits and confirmed their genetic correlations. With the aim of predicting microbiota-related host traits, chapter three examined the composition of the ileum microbiota and differential abundance analysis (DAA). Based on this study, it was shown that a sex-specific influence on microbiota composition exists. The digesta samples of all animals were dominated by five genera, which contributed to more than 70% of the total ileum microbial community. In examining the microbiota composition of each of the 50 animals with the highest and lowest PU, DAA revealed genera significantly associated with PU. In chapter four, we characterized the influence of performance-related gut microbiota to unravel the microbial architecture of the traits evaluated. The aim of this study was to determine whether the variation in PU is partly driven by the microbial community in the ileum. We used microbial mixed linear models to estimate microbiabilities (m^2). This determines the fraction of phenotypic variance that can be explained by the gut microbiota. The estimation of m^2 was 0.15 for PU and was highly significant. It was also highly significant for feed intake, body weight gain and feed per gain. This model was bivariately extended and showed a high microbial correlation of the traits. Based on both results, the ileum microbiota composition plays a substantial role in PU as well as in performance traits, and there is a considerable animal microbiota correlation, showing that the microbiota affects multiple traits. The microbial drivers of this microbial fraction were identified by applying microbiome-wide association studies (MWAS). By back-solving the microbial linear mixed model, we approximated the effect of single OTUs on the phenotypic traits from the microbial model solutions. An MWAS at the genus level uncovered several traits associated with bacterial genera. Subsequently, we assessed whether the microbial community in the ileum is a heritable host trait that can be used for breeding individuals with improved PU. In chapter five we applied QTL analysis using specific genera to examine whether they are linked with genomic SNP markers. These QTL analyses revealed a link between some microbiota species and host genomic regions of chromosomes and SNP markers. By estimating significant heritabilities for some genera, we were able to provide evidence for the hypothesis that the microbial community and microbial features are at least partially related to host genetics. We predicted the animal microbial effects on PU and correlated performance traits by applying microbial best linear unbiased predictions (M-BLUP). In addition, genomic best linear unbiased predictions (G-BLUP) were used to predict the SNP effect for the predicted animal microbial effect. A combination of those two may help to predict genomic breeding values of the microbiota effects for future hologenomic breeding programs.
Breeding objectives and selection schemes for Boran cattle in Kenya
(2009) Rewe, Thomas; Valle Zárate, Anne
Beef production with Boran cattle of Kenya presents an opportunity for utilising the potential of an indigenous breed. Improving the performance of these cattle through production of quality breeding stock may support the livelihoods of Boran cattle farmers. Organised breeding programmes for Boran cattle in Kenya are lacking. This thesis focussed on the design of optimal genetic improvement programmes for Boran cattle raised in the semi-arid tropics of Kenya. Specifically, the aims were: 1) to review the potential for beef cattle genetic improvement in sub-Saharan Africa, 2) to describe the state of institutional framework supporting Boran breeding in Kenya while considering the different categories of Boran cattle farmers, 3) to investigate the genetic and economic merit of alternative breeding programmes based on improved Boran, the unimproved Boran and the possibilities of expanding an inclusive breeding programme for these two strains of Boran cattle, 4) to discuss the feasibility of alternative breeding strategies within the context of a formal breeding programme in Kenya. The methodological approach entailed a review of the literature on cattle production and genetic improvement strategies for sub-Saharan Africa. This was done by identifying previous and on-going breeding activities among indigenous cattle breeds based on their regional distribution in Africa. An institutional framework analysis to characterise the Boran breeding sector in Kenya was also performed. Open questions were presented to the Boran cattle Breeders? Society of Kenya through an online survey to ascertain the structure of the society in terms of membership, cattle populations and animal breeding activities. The production systems, cattle types and breeding objectives were also sought from previous studies on Boran cattle in the process of developing appropriate breeding programmes. Design and evaluation of nucleus breeding programmes (genetically and economically) was done with the ZPLAN computer programme by defining the breeding objectives and selection criteria traits, describing breeding and commercial populations, describing selection groups as well as their reproduction performance parameters. The costs of the breeding programme included fixed costs and costs of animal recording. To account for genetic gain and the flow of animal genetics, a gene transmission matrix was defined utilising the selection groups alongside genetic and phenotypic parameter matrices. The information sources for the selection criteria were mainly parental selection groups and halfsibs of animal. The number of animals forming the selection groups and information sources was calculated in the NBILD and NUMBER subroutines of the ZPLAN. The interest rates for returns and costs were 8% and 6% respectively while the investment period was set at 25 years. The scope of the study was limited to two classes of farmers keeping Boran, the commercial beef ranchers and the market-oriented low-input beef producers that interact with commercial beef ranchers. Three breeding objectives were evaluated, 1) conventional breeding objectives with market (economic) values derived from bio-economic modelling, namely: direct sale weight, dressing percentage, consumable meat percentage, cow weaning rate, cow survival rate, cow weight, age at first calving, milk yield, feed intake and post weaning survival rate, 2) a combination of selected conventional target traits in addition to traits important to low-input farmers to exploit the ongoing informal interaction between the large scale ranchers and low-input systems, and 3) trait preferences for low-input farmers derived from conjoint analysis studies namely; sale weight, calving interval, temperament, tick resistance, trypanotolerance and lactation milk yield. To evaluate the benefit of perceived trypanotolerance in unimproved low-input herds, strategic recording for trypanotolerance for offspring of nucleus sires born in these herds was assumed. Closed and open-nucleus types were evaluated and variations on the nucleus size (5%, 10% and 25%). proportion of gene transfer to commercial herds (25%, 50%, 70%) and the proportion of gene importation into the nucleus (10%, 20% and 30%) were tested. The results from the institutional framework analysis showed that the Boran sector is structured with a section of the farmers being large scale commercial ranchers keeping approximately 17% of a total population of 580,000 heads of cattle. The rest were Boran farmers operating in low-input production systems keeping over 80% of the total population. The large scale commercial ranchers were found to be divided into two groups, about 52% of these farmers were elite breeders that record with the Kenya Stud Book and the rest were mainly commercial. The large scale commercial ranchers keep the improved Boran while the low-input farmers keep the unimproved Boran. The large scale commercial ranchers were organised into a breed society, namely, the Boran Cattle Breeders Society (BCBS), incorporating both the elite breeders and the commercial group. The BCBS was identified as a key stakeholder in the breeding of Boran cattle because of their informal role as suppliers of breeding stock. The results from the evaluation of alternative open and closed-nucleus breeding programmes utilising the Boran cattle populations were obtained with the ZPLAN computer programme. For the elite breeders?, where a total population of 52,000 cows with a breeding unit of 25% was assumed, the overall monetary genetic gain was KSh86 per cow while the profit per cow was KSh361 under the conventional breeding objective. The breeding programme with the entire BCBS group where a population of 99,972 cows was assumed obtained a higher monetary genetic gain and profit than the elite group per cow of KSh93 and KSh431 respectively under the same breeding objective. The results revealed the effect of a larger effective population size on performance of breeding programmes. The breeding programme based wholly on market oriented low-input producers was evaluated using farmer trait preferences as the breeding objective. This breeding programme posted a negative gain for milk yield of -1.1 kg, which improved when restrictions on growth and adaptation were applied. The introduction of the combined breeding objective that included adaptation and disease tolerance traits resulted in a drop in sale weight gain by almost 2 kg. However, post-weaning survival rate improved from 0.4% to 1% and trypanotolerance gained 20% packed cell volume within this breeding objective. There was reduction in feed intake under the combined breeding objective, which is desirable considering the prevailing limitations on land, feed and climatic conditions. This may induce a change in focus from the continuous improvement in sale weight. The gains in post weaning survival rate would support this objective. The results from the expanded breeding programme may be beneficial to both the low-input farmers and the commercial ranchers because of the advantages conferred from the improvement in adaptation traits. The benefits of extra recording for trypanotolerance in the commercial herds of the expanded programme were not realised. In general, the open-nucleus programmes were superior genetically while the closed-nucleus programmes were superior economically. The larger nucleus sizes (25%), higher gene contributions to commercial herd (70%) and limiting nucleus opening to 10% were most profitable. The limitations of the study were observed from the online interviews with respect to the amount of information that could be retrieved from key persons. Similarly, information on the legal framework of the breeding sector was scarce since Kenya has no active livestock breeding policy. The design and evaluation of the breeding programmes was possible with ZPLAN, however, in this study, genetic variance for traits, which normally diminishes with selection and inbreeding, was not account for. This may have had implications related to overestimation of genetic response and economic returns. Nonetheless, the potential of the Boran for both beef production and fitness traits coupled with the presence of institutional support for animal recording in Kenya were evaluated as strengths of the system. This study has shown the possibilities of combining market and non-market traits useful in breeding programmes for cattle utilised in different production systems. This approach is useful in cases where interactions exist between different categories of farmers. To benefit from advantages offered by open-nucleus breeding, recording may be avoided in the commercial herds and selection be done under criteria that are acceptable by the farmers. Further investigations on farmer organisations and comprehensive livestock breeding policies may aid the process of establishing co-ordinated breeding programmes for Boran cattle in Kenya.
Genetic analyses of feather pecking and related behavior traits of laying hens
(2016) Lutz, Vanessa; Bennewitz, Jörn
The main objective of the present study was to study the genetic foundation of behaviour traits, especially feather pecking behaviour, and to infer ethological interrelationship between certain traits of laying hens. The data of two divergently selected lines for feather pecking behaviour was available, and additionally a large F2-cross, set up from these divergently selected lines, was established. Chickens of a White Leghorn layer line were divergently selected for high and low feather pecking for 11 generations. The selection started in the Danish Institute of Animal Sciences, Foulum, Denmark, for the first six generations (0-5). Thereafter, five rounds of selection took place at the Institute of Animal Science, University of Hohenheim, Germany. The large F2-cross was established from the 10th selection generation, and a comprehensive data collection of behaviour and performance traits of 960 hens was performed. These two data sets were used for the following five research chapters. In chapter one, a quantitative genetic analysis of fear traits and feather pecking as well as aggressive pecking using data from the large F2-cross was performed. Fear was recorded by the tonic immobility test, the open field activity and the emergence box test. These were recorded at a juvenile and adult age. Behavior traits as feather pecking and aggressive pecking were recorded in groups of 36 to 40 animals at the age of 27 weeks. The genetic parameters were estimated using a linear mixed model. Aggressive pecking showed the highest heritability (0.27) followed by feather pecking (0.14). The fear test traits showed heritabilities in the range of 0.07 to 0.14. The appreciable genetic correlation between fear traits and feather pecking was tonic immobility at juvenile age (rg=0.27). In chapter two we used dispersed Poisson models to estimate variance components, heritabilities of feather and aggressive pecking of different observation periods. The short period included the number of feather pecks in 20 min and the medium period was the summed bouts within one day. The results showed that modelling the data as repeated observations (short and medium period) and analysing them with a dispersed Poisson model is a suitable option to separate the important permanent environment effects from the additive animal effects and to account for the non-normal distribution of the data. The objective of chapter three was to analyze the interrelationship between feather pecking and feather eating as well as general locomotor activity using structural equation models. The estimated heritabilities of feather eating, general locomotor activity and feather pecking were 0.36, 0.29 and 0.20, respectively. The genetic correlation between feather pecking and feather eating (general locomotor activity) was 0.17 (0.04). A high genetic correlation of 0.47 was estimated between feather eating and general locomotor activity. The recursive effect from feather eating to feather pecking was λ ̂_(FP,FE)= 0.258, and from general locomotor activity to feather pecking λ ̂_(FP,GLA)= 0.046. These results imply that an increase of feather eating leads to an increased feather pecking behavior and that an increase in general locomotor activity results in a higher feather pecking value. The objective of chapter four was to perform a quantitative genetic analysis and to map signatures of selection in two divergent laying hen lines selected for feather pecking behaviour. In the selection experiment, lines were selected for low or high feather pecking for 11 generations. Pedigree and phenotypic data were available for the last six generations of both lines for the statistical analysis with a standard mixed linear model and a Poisson model. The mixed linear model failed to analyse the low feather pecker data because of the large number of 0s in the observation vector. The Poisson model fitted the data well and revealed a small but continuous genetic trend in both lines. From the 11th generation 75 birds, 41 high feather peckers and 34 low feather peckers were genotyped using the Illumina 60K chicken Infinium iSelect chip. An FST-based approach was used to map selection signature. We detected 17 genome-wide significant SNPs with a FST-value of 1, i.e. alleles were divergently fixed in the two lines, which are mostly located on chromosome 3 and 4, and a number of additional significant SNPs with a p-value of ≤ 5x10-4 and ≤ 5x10-5, respectively. Based on the assumption that selection affects several consecutive SNPs, 13 clusters were identified. In chapter five, we used the data from the large F2-cross experiment to perform a genome-wide association study for feather pecking and aggressive pecking behaviour, to combine the results of this GWAS with the results from the selection experiment (chapter four) in a meta-analysis, and to link the results to those obtained from a differential gene expression study. 817 F2-hens were genotyped with the Illumina 60K chicken Infinium iSelect chip. We used single marker association analysis and a Poisson model. We detected four genome-wide significant SNPs for aggressive pecking delivered, but none for feather pecking and aggressive pecking received. However, a number of significant SNPs at p≤5x10-5 were mapped for feather pecking and aggressive pecking received. In the meta analysis we identified nine genome-wide significant SNPs for feather pecking delivered, which were localized in chromosomal clusters (3 Mb). A previously conducted differential gene expression analysis provided eight significantly differential expressed genes within the feather pecking associated chromosomal clusters. The thesis ends with a general discussion.
Genome-wide mapping and functional analysis of genes determining the meat quality in pigs
(2014) Stratz, Patrick; Bennewitz, Jörn
In chapter one QTL were mapped and tested for pairwise epistasis for meat quality traits in three connected porcine F2 crosses comprising around 1000 individuals. The crosses were derived from Chinese Meishan, European Wild Boar and Piétrain. The animals were genotyped genomewide for approximately 250 genetic markers and phenotyped for seven meat quality traits. QTL mapping was done using a multi-QTL multi-allele model. It considered additive (a), dominance (d) and imprinting (i) effects. The major gene RYR1:G.1843C>T affecting the meat quality was included as a cofactor in the model. The mapped QTL were tested for possible epistatic effects between the main effects, leading to nine orthogonal forms of epistasis (aa, ad, da, di, id, ai, ia, dd and ii). Numerous QTL were found; the most interesting are located on chromosome SSC6. Epistasis was significant (FDR q-value<0.2) for the pairwise QTL on SSC12 and SSC14 for pH 24 h after slaughter and for the QTL on SSC2 and SSC5 for rigour. In chapter two around 500 progeny tested Piétrain sires were genotyped with the PorcineSNP60 BeadChip. After data filtering around 48k SNPs were useable in this sample. These SNPs were used to conduct a genome-wide association analysis for growth, muscularity and meat quality traits. Because it is known, that a mutation in the RYR1 gene located on chromosome 6 shows a major effect on meat quality, this mutation was included in the models. Single-marker and multi-marker association analysis were performed. The results revealed between one and eight significant associations per trait with P-value<0.00005. Of special interest are SNPs located on SSC6, 10 and 15. In chapter three a literature search was conducted to search putative candidate genes in the vicinity of significant SNPs found in the association analysis. MYOD1 was suggested as putative candidate gene. The expression of MYOD1 was measured in muscle tissue from 20 Piétrain sires. Growth, muscularity and meat quality traits were available. DNA was isolated out of blood tissue to genotype the SNP ASGA0010149:g. 47980126G>A. Significant Correlations (FDR q-value<0.15) between the expression of MYOD1 and growth and muscularity traits were found. Association between the traits, respectively MYOD1, and ASGA0010149:g. 47980126G>A was tested, but was only significant (FDR q-value<0.15) for two muscularity traits. In chapter four the LD structure in the genome of the Piétrain pigs was characterized using data from the PorcineSNP60 BeadChip. The Relative Extended Haplotype Homozygosity test was conducted genome-wide to search for selection signatures using core haplotypes above a frequency of 0.25. The test was also conduct in targeted regions, where significant SNPs were already found in association analysis. A small subdivision of the population with regard to the geographical origin of the individuals was observed. As a measure of the extent of linkage disequilibrium, r2 was calculated genome-wide for SNP pairs with a distance 5Mb and was on average 0.34. Six selection signatures having a P-value<0.001 were genome-wide detected, located on SSC1, 2, 6 and 17. In targeted regions, it was possible to successfully annotate nine SNPs to core regions. Strong evidence for recent selection was not found in those regions. Three selection signatures with P-value<0.1 were detected on SSC2, 5 and 16. To reduce the costs of genomic selection, selection candidates can be genotyped with an SNP panel of reduced density (384 SNPs). The aim of chapter five was to investigate two strategies for the selection of SNPs to be considered in the above mentioned SNP panel, using 895 progeny tested and genotyped German Piétrain boars. In the first strategy equal spaced SNPs were selected, which were used to impute the high density genotypes. In the second strategy SNPs were selected based on results of association analysis. Direct genomic values were estimated with GBLUP from deregressed estimated breeding values. Accuracies of direct genomic values for the two strategies were obtained from cross validation. A regression approach to correct for the upward bias of the cross validation accuracy of the direct genomic values was used. The first strategy resulted in more accurate direct genomic values. This implies that imputation is beneficial even if only 384 SNPs are genotyped for the selection candidates.
Genomic analyses of behavior traits in laying hen lines divergently selected for feather pecking
(2021) Iffland, Hanna; Bennewitz, Jörn
Feather pecking is a longstanding problem in commercial layer flocks. It often causes injured birds and even cannibalism. In the past, hens were beak trimmed to reduce feather pecking. Nevertheless, this procedure is already prohibited in some EU countries. Hence, a solution to this problem is urgently needed. The experimental populations analyzed in this thesis were formed by hens based on a White Leghorn layer strain which were divergently selected for high and low feather pecking since 1995. The first experimental population of this thesis was an F2 cross of about 900 hens which was established of the 10th generation of the pure selection lines. The second population consisted of about 500 hens of the 15th generation of these two lines. The aim of this thesis was to gain further knowledge of the genetic background of feather pecking and its relation to additional behavior traits and the gut microbiome. In chapter one, a novel model to detect extreme feather pecking hens was developed. Therefore, a mixture of two negative binomial distributions was fitted to feather pecking data of the F2 cross. With the estimated parameters, the trait posterior probability of a hen to belong to the extreme feather pecking subgroup (pEFP) was calculated. The fear tests tonic immobility and emerge box were conducted at juvenile and adult age of the hens to relate fearfulness to pEFP. After dichotomization, all traits were analyzed in a multivariate threshold model and subsequent genomewide association studies (GWAS) were performed. The fit revealed that extreme feather peckers made up a proportion of about one third of the hens. The new trait pEFP has a medium heritability of 0.35 and is positively correlated with the fear traits. Breeding for this new trait could be an option to reduce the proportion of extreme feather peckers. An index of fear related traits might serve as a proxy to breed indirectly against pEFP. In chapter two, the model to detect extreme feather pecking hens was applied to the pure selection lines. After calculation of the trait pEFP, GWAS with a subsequent post GWAS analysis were performed. Additionally, to find genomic regions influencing feather pecking, selection signatures were mapped by applying the intra-population iHS and the inter-population FST approach. Mapping of selection signatures revealed no clear regions under selection. GWAS revealed a region on chromosome one, where the existence of a quantitative trait locus (QTL) influencing feather pecking is likely. The candidate genes found in this region are a part of the GABAergic system. Despite the polygenic nature of feather pecking, selection on these candidate genes may reduce the extreme occurrence of it. In chapter three, the relation between agonistic behavior and feather pecking was analyzed. Therefore, the active parts of the traits (delivery of feather pecking, aggressive pecking or threatening) as well as the passive parts (reception of the traits) were considered. These groups of traits were additionally summarized by means of an index formation which led to the two additional traits Activity and Passivity, because all these behaviors are undesired in their excessive manifestations. Moreover, Indices were built by subtracting the passive traits from the respective active traits to obtain the feather pecking index, the aggression index and the threat index. Phenotypic correlations were estimated between all traits which were followed by heritability estimations and GWAS. Feather pecking is significantly positively correlated with the agonistic traits in both lines. The active traits and the feather pecking index show medium heritabilities. Hence, selection on high feather pecking leads to an increase of agonistic behavior whereas the correlation probably depends on the phase of establishing the social hierarchy and might disappear, after a stable ranking is established. GWAS revealed that the heritable traits in this study seem to be typical quantitative traits. Chapter four provides the analyses of the gut microbial composition of the two feather pecking lines, followed by the estimation of microbiabilities for feather pecking and the two agonistic behavior traits, to study the influence of the gut microbiome on behavior. Microbiota samples from digesta and mucosa were taken from ileum and caecum. The microbial communities were determined by using 16S RNA gene sequencing techniques. Although both lines differ significantly in some fractions of their gut microbial composition, the microbial animal effects were mostly negligibly small. Thus, the calculated microbiabilities were close to zero and not significant in both lines and for all traits investigated. Hence, trait variations were not affected by the gut microbial composition in both feather pecking lines. The thesis ends with a general discussion where additional results of a meta-analysis of pEFP and breeding strategies against feather pecking are considered.
Genomic and microbial analyses of quantitative traits in poultry
(2023) Haas, Valentin Peter; Bennewitz, Jörn
Feed and nutrient efficiency will become increasingly important in poultry production in the coming years. In addition to feed efficiency, particular attention is paid to phosphorus (P) in nonruminants. Especially growing animals have a high demand of P but through the low usability of plant-based P sources for nonruminants, mineral P is added to their feeds. Due to worldwide limited mineral P sources, the high environmental impact of P in excretions and high supplementation costs, a better utilization of P from feed components is required. Animals’ P utilization (PU) is known to be influenced by the host genetics and by gastrointestinal microbiota. The overall aim of this thesis was to investigate the relationships between host genetics, gastrointestinal microbiota composition and quantitative traits with the focus on PU and related traits in F2 cross Japanese quail (Coturnix japonica). Japanese quail represent a model species for agriculturally important poultry species. In Chapter one, a genetic linkage map for 4k genome-wide distributed SNPs in the study design was constructed and quantitative trait loci (QTL) linkage mapping for performance as well as bone ash traits using a multi-marker regression approach was conducted. Several genome-wide significant QTL were mapped, and subsequent single marker association analyses were performed to find trait associated marker within the significant QTL regions. The analyses revealed a polygenic nature of the traits with few significant QTL and many undetectable QTL. Some overlapping QTL regions for different traits were found, which agreed with the genetic correlations between the traits. Potential candidate genes within the discovered QTL regions were identified and discussed. Chapter two provided a new perspective on utilization and efficiency traits by incorporating gastrointestinal microbiota and investigated the links between host genetics, gastrointestinal microbiota and quantitative traits. We demonstrated the host genetic influences on parts of the microbial colonization localized in the ileum by estimating heritabilities and mapping QTL regions. From 59 bacterial genera, 24 showed a significant heritability and six genome-wide significant QTL were found. Structural equation models (SEM) were applied to determine causal relationships between the heritable part of the microbiota and efficiency traits. Furthermore, accuracies of different microbial and genomic trait predictions were compared and a hologenomic selection approach was investigated based on the host genome and the heritable part of the ileum microbiota composition. This chapter confirmed the indirect influence of host genetics via the microbiota composition on the quantitative traits. Chapter three further extended the approaches to identify causalities from chapter two. Bayesian learning algorithms were used to discover causal networks. In this approach, microbial diversity was considered as an additional quantitative trait and analyzed jointly with the efficiency traits in order to model and identify their directional relationships. The detected directional relationships were confirmed using SEM and extended to SEM association analyses to separate total SNP effects on a trait into direct or indirect SNP effects mediated by upstream traits. This chapter showed that up to one half of the total SNP effects on a trait are composed of indirect SNP effects via mediating traits. A method for detecting causal relationships between microbial and efficiency traits was established, allowing separation of direct and indirect SNP effects. Chapter four includes an invited review on the major genetic-statistical studies involving the gut microbiota information of nonruminants. The review discussed the analyses conducted in chapter one to three and places the analyses published in these chapters in the context of other statistical approaches. Chapter four completed the microbial genetic approaches published to date and discussed the potential use of microbial information in poultry and pig breeding. The general discussion includes further results not presented in any of the chapters and discusses the general findings across the chapters.
Genomische und mikrobielle Analysen von Effizienzmerkmalen beim Schwein
(2022) Weishaar, Ramona Ribanna; Bennewitz, Jörn
Most traits in animal breeding, including efficiency traits in pigs, are influenced by many genes with small effect and have moderate heritabilities between 0.1 and 0.5, which enables efficient selection. These so-called quantitative traits are influenced by genetic factors and environmental factors. The use of next-generation sequencing methods, such as 16S rRNA sequencing to analyse the gut microbiome of livestock, allows identification and analysis of the gut microbiota. It has been shown that the composition of the microbiota in the gastrointestinal tract is heritable and has an influence on efficiency traits. Thus, the animal genome influences the phenotype not only directly by altering metabolic pathways, but also indirectly by changing the composition of the microbiota. This increases the interest in implementing gut microbiota into existing breeding strategies as an explanatory variable. The potential of an efficient utilization and absorption of nutrients varies between individuals. Differences in nutrient absorption depend on feed intake, digestion of dietary components in the stomach and intestine, and intake of digested nutrients from the gastrointestinal tract into blood and lymphatic vessels. Undigested nitrogen is excreted as urea and can be detected by blood urea nitrogen (BUN). The BUN is correlated with efficiency traits and there exist differences between pig breeds. Thus, therefore the BUN would be conceivable as an easier recordable trait for nitrogen utilisation efficiency in pig breeding. In the first chapter of this study, an existing data set of the Department for Animal Genetics and Breeding of the University of Hohenheim was used. This is a data set with 207 phenotyped and genotyped Piétrain sows. The relationship between gut microbial composition, efficiency traits and the porcine genome is investigated using quantitative genetic methods. The heritabilities of the traits FVW, RFI, TZ, and FI ranged from 0.11 to 0.47. The microbiabilities of the traits were significant and ranged from 0.16 to 0.45. In a further step, the previously generated microbial animal effects were used as observation vector for a genomic mixed model. Subsequently, heritabilities for the microbial animal effect were estimated, ranging from 0.20 to 0.61. The similarity of the heritabilities and microbiabilities suggests that the traits are influenced to a similar extent by both genetics and gut microbiota and that the microbial animal effect is determined by the host. These results are underlined by the identification of genera and phyla with significant effects on efficiency traits. The microbial architecture of the traits demonstrated a poly-microbial nature, there are many OTUs with small effects involved in the variation of the observed traits. Genomic Best Linear Unbiased Predictions (G-BLUP) and Microbial Best Linear Unbiased Predictions (M-BLUP) were performed to predict complex traits. The accuracies of M-BLUP and G-BLUP were all in a similar range between 0.14-0.41. This shows that gut microbiota could be used to predict performance traits or be included as a variable in the existing models of breeding value estimation to realize an increase in accuracies. The second part of the paper analysed a dataset from a research project called "ProtiPig". The data set included 475 sows and castrates of crossbreds of German Landrace x Piétrain and was analysed for protein utilization efficiency and nitrogen(N)-utilization efficiency. N-utilization efficiency is a trait that is difficult to record. Because conventional metabolic cage methods are a very complex procedure and difficult to integrate in the standard recording, it was tested whether the BUN is suitable as a proxy trait. Moderate to medium heritabilities could be estimated for all traits and ranged from 0.13 to 0.49. The genome-wide association studies showed that the traits were polygenic. For the BUN, SNPs could be detected that were above the genome-wide significance level. Significant genetic and phenotypic correlations were found between some traits. In particular, the heritabilities of BUNs and the significant genetic correlation between BUN and N-utilization efficiency indicate an opportunity to use the BUN to select for improved N-utilization efficiency. Before the research results generated here can be implemented in breeding practice, further questions must be clarified. In addition, a larger number of animals is needed to validate the results. The results presented here demonstrate the potential of microbial-assisted breeding value estimation and the use of BUN to identify selection candidates for breeding for increased efficiency.
Investigations on major gene by polygene and gene by environment interaction in German Holstein dairy cattle
(2014) Streit, Melanie; Bennewitz, Jörn
Putative interaction effects between DGAT1 K232A mutation and the polygenic terms (all genes except DGAT1) were investigated in chapter one. This was done for five milk production traits (milk yield, protein yield, fat yield, protein percentage and fat percentage) in the German Holstein dairy cattle population. Therefore, mixed models are used. The test for interaction relied on the comparison of polygenic variance components depending on the sire?s genotypes at DGAT1 K232A. Found substitution effects were highly significant for all traits. Significant interactions between DGAT1 K232A and the polygenic term were found for milk fat and protein percentage. These interactions could be used in breeding schemes. Depending on the DGAT1 K232A genotypes of the sample, in which the sire will be used, three polygenic breeding values of a sire can be calculated. Because the genotypes of the samples are often unknown and usually heterogeneous, this is not a practical approach. Rank correlations between the three polygenic EBVs were always above 0.95, which suggested very little re-ranking. GxE were studied in chapter two. For this, reaction norm random regression sire models were used in the German Holstein dairy cattle population. Around 2300 sires with a minimum of 50 daughters per sire and at minimum seven first-lactation test day observations per daughter were analyzed. As traits, corrected test day records for milk yield, protein yield, fat yield and somatic cell score (SCS) were used. As environmental descriptors, we used herd test day solutions (htds) for milk traits, milk energy yield or SCS. Second-order orthogonal polynomial regressions were applied to the sire effects. Results showed significant slope variances of the reaction norms, which caused a non-constant additive genetic variance across the environmental ranges considered, which pointed to the presence of minor GxE effects. When the environment improved, the additive genetic variance increased, meaning higher (lower) htds for milk traits (SCS). This was also influenced by pure scaling effects, because the non-genetic variance increased in an improved environment and the heritability was less influenced by the environment. For the environmental ranges considered in this study, GxE effects caused very little re-ranking of the sires. To obtain unbiased genetic parameters, it was important to model heterogeneous residual variances. A large genome-wide association analysis was conducted in chapter three to identify SNPs that affect general production (GP) and environmental sensitivity (ES) of milk traits. Around 13 million daughter records were used to calculate sire estimates for GP and ES with help of linear reaction norm models. Daughters were offspring from 2297 sires. The sires were genotyped with a 54k SNP chip. As environmental descriptor, the average milk energy yield performance of the herds at the time where the daughter observations were recorded was used. The sire estimates were used as observations in genome-wide association analyses using 1797 sires. With help of an independent validation set (500 sires of the same population), significant SNPs were confirmed. To separate GxE scaling and other GxE effects, the observations were log-transformed. GxE effects could be found with help of reaction norm models and numerous significant SNPs could be validated for GP and ES, whereas many SNPs affecting GP also affected ES. ES of milk traits is a typical quantitative trait, which is controlled by many genes with small effects and few genes with larger effect. Effects of some SNPs for ES were not removable by log-transformation of observations, indicating that these are not solely scaling effects. Positions of founded clusters were often in well-known candidate regions affecting milk traits. No SNPs, which show effects for GP and ES in opposite directions could be found. Environmental descriptor in GxE analyses is often modelled by average herd milk production levels. Another possibility could be the level of hygiene and udder health. In chapter four, the same models were used as in chapter three. A genome-wide association analysis was done using htds for SCS as an environmental descriptor. With help of this, several SNP clusters affecting intercept and slope could be detected and confirmed. Many SNPs or clusters affecting intercept and slope could be identified, but in total, the number of SNPs affecting intercept was larger. The same SNPs could be detected and validated with and without considering GxE in reaction norm models. Some SNPs affecting only slope were found. For slope, nearly the same SNPs could be found with SCS as an environmental descriptor as presented in chapter three, although both environmental descriptors were only slightly correlated.
Investigations on methodological and strategic aspects of genomic selection in dairy cattle using real and simulated data
(2018) Plieschke, Laura Isabel; Bennewitz, Jörn
In Chapter one a method was developed to separate the genomic relationship matrix into two independent covariance matrices. Here, the base group component describes the covariance that results from systematic differences in allele frequencies between groups at the pedigree base. The remaining segregation component describes the genomic relationship that is corrected for the differences between base populations. To investigate the proposed decomposition three different models were tested on six traits, where the covariance between animals was described either only by the segregation component or by a combination of the two components. An additional variant examining the effect of a fixed modeling of the group effects was included. In total, 7965 genotyped Fleckvieh and 4257 genotyped Brown Swiss and 143 genotyped Original Braunvieh bulls were available for this study. The proposed decomposition of the genomic relationship matrix helped to examine the relative importance of the effects of base groups and segregation component in a given population. It was possible to estimate significant differences between the means of base groups in most cases for both breeds and for the traits analyzed. Analysis of the matrix of base group contributions to the populations investigated revealed several general breed-specific aspects. Comparing the three models, it was concluded that the segregation component is not sufficient to describe the covariance completely. However, it also was found that the model applied has no strong impact on predictive power if the animals used for validation show no differences in their genetic composition with respect to the base groups and if the majority of them have complete pedigrees of sufficient depth. The subject of the chapter two was investigation to systematically increase the reliability of genomic breeding values by integrating cows into the reference population of genomic breeding value estimation. For this purpose a dataset was generated by simulation resembling the German-Austrian dual-purpose Fleckvieh population.. The concept investigated is based on genotyping a fixed number of daughters of each AI bull of the last or last two generation of the reference population and, together with their phenotypic performance, to integrate them into the reference population of the genomic evaluation. Different scenarios with different numbers of daughters per bull were compared. In the base scenario the reference population was made up of 4200 bulls. In the extended scenarios, more and more daughters were gradually integrated in the reference population. The reference population of the most extended scenario contained 4200 bulls and 420,000 cows. It was found that the inclusion of genotypes and phenotypes of female animals can increase the reliabilities genomic breeding values considerably. Changes in validation reliability of 6-54% for a trait with a heritability of 0.4 depending on scenario were found. As the number of daughters increased, the validation reliability increased as well. It should be noted that the composition of the daughter samples had a very great influence on whether the additional genotyped and phenotyped animals in the reference population can have a positive effect on the reliability of genomic breeding values. If pre-selected daughter samples were genotyped, the mean validation reliability decreased significantly compared to a correspondingly large unselected daughter sample. In addition, a higher bias was observable in these cases. Chapter three expands the investigations of chapter two by a low-heritability trait, as well as the aspect of so called new traits. The results found in chapter two were confirmed in chapter three for a low-heritability trait. Changes in validation reliability of 5-21% for a heritability of 0.05 depending on scenario were found. The negative effects of pre-selected daughter samples were even more pronounced in chapter three. In the case of an ‘old’ trait, the number of phenotypes is expected to be (nearly) unlimited, since a recording system is well established. In the case of a new trait recording of phenotypes just started, therefore the number of phenotypes is limited. Two different genotyping strategies were compared for new traits. On the one hand, the sires of the phenotyped cows were genotyped and on the other hand the phenotyped cows were genotyped themselves. It was found in all compared scenarios that it is more sensible to genotype cows themselves instead of the genotyping their sires. However, if usual strategy of phenotyping female animals and genotyping of males is applied, it is at least important to ensure that many daughters are phenotyped in a balanced system. If different numbers of daughters per bull are phenotyped and unbalancedness becomes severe, the average validation reliability decreased significantly.
Pedigreeanalysen zur Beschreibung der populations- und quantitativgenetischen Situation von baden-württembergischen Lokalrinderrassen
(2014) Hartwig, Sonja; Bennewitz, Jörn
The challenge of a conservation breeding program is to solve a conflict of goals: genetic variability and genetic autonomy should be maintained, and on time a certain amount of breeding progress has to be realized to ensure the ability to compete. The aim of the present study was to analyse the situation concerning the targets mentioned above for traditional cattle breeds of Baden-Württemberg. Furthermore, methods for the consolidation and development of these breeds should be reconsidered. In chapter 1 the organisation of a breeding program for small cattle breeds is discussed. The challenge of such a program is the conservation of agrobiodiversity, culture and traditions and the progress of traditional local breeds. Number and extend of these breeds declined due to the increasing popularity of high-yielding breeds. Additionally, some of the local breeds are used in different branches of production compared to their original usage. Breeding programs have to be fitted. The establishment of individual adapted breeding methods is required for a sophisticated solution of the conflict mentioned above. Federal support is requested. Nowadays the implementation of genomic selection is not yet practicable for small breeds. But future opportunities should be analysed. The establishment of performance tests concerning breed specific product and efforts is demanded to improve these characteristics. In the following genetic variability of Vorderwald, Hinterwald and Limpurg cattle was examined. In chapter 2 for each breed two reference populations were defined that enable to observe the development over the years. Animals within the reference population comply with restrictions concerning racial origin and completeness of pedigree. Effective population size and the effective number of founders, and ancestors were estimated. The interpretation of the results was done with regard to the history of the breeds. The absolute population size of Vorderwald cattle is the biggest one. The number of Hinterwald cattle is intermediate; Limpurg cattle have the smallest population size. Whereas the management of Hinterwald cattle seemed to maintain genetic variability, the management of Vorderwald cattle was not that target-orientated. With an effective population size greater than 100 there is enough genetic variability within Vorderwald and Hinterwald. In contrast the values of Limpurg cattle are too low. Besides genetic variability, genetic autonomy and breeding progress are the targets of a conservation breeding program. Cross-breeding was carried out to mitigate the risk of inbreeding depression and to improve the performance of local breeds. However, breeding activities for local breeds were not as intensive and target oriented as for popular high yielding breeds. While the gap between the performance of high-yielding and local breeds increase, genetic autonomy of local breeds declined due to migrant influences. Chapter 3 examined the importance of migrant breed influences for the realization of breeding progress of beef traits of Vorderwald and Hinterwald cattle. The results show that there is a high amount of migrant contributions and their effects on performance are substantial for most traits. Breeding values adjusted for the effects of the migrant breeds showed little genetic trend for beef traits. The subject of chapter 4 is the development of milk yield and the associated migrant influences. Yield deviations for milk, fat, and protein content were analysed. Migrant contributions to Vorderwald cattle were high and even rose in the latest past. All the effects of foreign breeds were positive and in most cases highly significant. Most influential breed was Montbéliard. The influences of migrant breeds were substantial for the development of milk performance. However, the trend of milk yield traits for both breeds was positive even without foreign breeds’ influences. In comparison the number of analysed Hinterwald cows was small due to the reason that just a few Hinterwald husbandries take part at the official milk performance recording. Migrant breed contributions were moderate and nearly constant over the time. The most influential migrant breed was the Vorderwald cattle. The development of milk yield shows a flat trend. Influences of migrant breeds were low. The thesis ends with a general discussion.
Selection methods for local breeds with historical introgression
(2018) Wang, Yu; Bennewitz, Jörn
For the management of local breeds with historical introgression, both genetic gain and the long-term evolution of genetic variability have to be taken into consideration. Traditional optimum contribution selection (traditional OCS) aims at maximizing genetic gain while controlling the rate of inbreeding by optimizing the genetic contribution of each selection candidate to the next generation. It is also a promising approach to maintain genetic diversity since the average kinship of selection candidates is restricted. However, for the breeds with historical introgression, this diversity may be caused by introducing genetic material from other breeds, which can be a risk of the conservation of small local populations. Therefore, the breeding objectives should not only focus on increasing genetic gain but also on maintaining the diversity of native alleles. The main aim of this project was to resolve the existing conflicts in the current breeding program of local breeds with historical introgression. Chapter 1 gave a brief introduction and background of the topic and formulated the objective of the thesis. In chapter 2, the current inbreeding status of German Angler cattle was evaluated based on both pedigree (F_PED) and genomic information. The genomic inbreeding coefficients of 182 Angler cattle were estimated via analyzing the genome proportion of run of homozygosity (F_ROH) and using the genomic relationship matrix (F_GRM). On average, the inbreeding level of Angler is relatively low compared to the other breeds ((F_PED ) ̅:0.013;(F_GRM ) ̅:-0.015; (F_(ROH>1Mb) ) ̅:0.031). Moderate to strong correlations (0.607–0.702) were found between F_PED and F_ROH based on different length categories of ROH segments. Moreover, it proved that F_ROH is a robust estimating method owing to its ability to capture both ancient and recent inbreeding. Although traditional OCS may achieve higher genetic gain with the restriction of the defined rate of inbreeding, in this case, inbreeding is not the main problem in the current breeding program and the advantage of OCS may be limited since the level of inbreeding may be lower than the threshold. In chapter 3, we developed the advanced optimum contribution selection strategy by considering migrant contribution and conditional kinship at native alleles in the OCS procedure. Different scenarios were compared for both functions of production and conservation based on pedigree information. It has been proved that the advanced OCS approach can effectively maintain the diversity of native alleles and genetic originality while ensuring genetic improvement with appropriate settings of constraint values. The availability of high-density single-nucleotide polymorphism (SNP) markers provides a solution for achieving accurate estimates of both coancestry and breed composition. In chapter 4 and chapter 5, we evaluated the long-term performance of advanced OCS strategies in both production and conservation function via simulating several subsequent generations based on genomic information. In chapter 4, we found that traditional OCS procedure has slight advantages in increasing genetic gain whilst controlling relatedness compared to truncation selection. However, the introgression of foreign genetic material by traditional OCS is not desirable in the local breed conservation. In the long run, constraining migrant contribution and kinship at native alleles in the OCS procedure is a promising approach to increase genetic gain whilst maintaining genetic uniqueness and diversity. Chapter 5 mimics a conservation program which aims at increasing the value of a breed for conservation by removing exogenous genetic material, maintaining within-breed genetic diversity, and increasing the genetic diversity among breeds. Simply minimizing the exogenous genetic contribution leads to the loss of both within and between population diversity. Moreover, the recovery process ended at a plateau after several generations. The best scenario was able to increase the native contribution from 0.317 to 0.706 before a segment-based kinship level of 0.10 was reached. This scenario maximized the native contribution, constrained the increase in kinship, and the increase in kinship at native alleles. Moreover, it constrained the mean kinship in a multi-breed core set to the current level, which is desirable for the conservation program. This thesis ends with a general discussion.
Using genome-wide association studies to map genes for complex traits in porcine F2 crosses
(2018) Schmid, Markus; Bennewitz, Jörn
In the era of genomics, genome-wide association studies (GWASs) have become the method of choice for gene mapping. This is still of great interest to infer the genetic architecture of quantitative traits and to improve genomic selection in animal breeding. Formerly, linkage analyses were conducted in order to map genes. Therefore, many F2 cross populations were generated by crossing genetically divergent lineages in order to create informative experimental populations. However, a small number of markers and the limited meiotic divisions led to imprecise mapping results. The main objective of the present study was to investigate the use of existing porcine F2 cross data, extended towards single nucleotide polymorphism (SNP) chip genotype information, for quantitative trait loci (QTL) mapping in the genomic era. A special focus was on mapping genes that also segregate within the Piétrain breed since this is an important sire line and genomic selection is applied in this breed. Chapter 1 is a review article of statistical models and experimental populations applied in GWASs. This chapter gives an overview of methods to conduct GWASs using single-marker models and multi-marker models. Further, approaches taking non-additive genetic effects or genotype-by-environment interactions into account are described. Finally, post-GWAS analysis possibilities and GWAS mapping populations are discussed. In chapter 2, the power and precision of GWASs in different F2 populations and a segregating population was investigated using simulated whole-genome sequence data. Further, the effect of pooling data was determined. GWASs were conducted for simulated traits with a heritability of 0.5 in F2 populations derived from closely and distantly related simulated founder breeds, their pooled datasets, and a sample of the common maternal founder breed. The study showed that the mapping power was high (low) in F2 crosses derived from distantly (closely) related founder breeds and highest when several F2 datasets were pooled. By contrast, a low precision was observed in the cross with distantly related founder breeds and the pooling of data led to a precision that was between the two crosses. For genes that also segregated within the common founder breed, the precision was generally elevated and, at equal sample size, the power to map QTL was even higher in F2 crosses derived from closely related founder breeds compared with the founder breed itself. Within and across linkage disequilibrium (LD) structures of such F2 populations were examined in chapter 3 by separately and jointly (pooled dataset) analyzing four F2 datasets generated from different founder breeds. All individuals were genotyped with a 62k SNP chip. The LD decay was faster in crosses derived from closely related founder breeds compared with crosses from phylogenetically distantly related founder populations and fastest when the data of all crosses were pooled. The pooled dataset was also used to map QTL for the economically important traits dressing out and conductivity applying single-marker and Bayesian multi-marker regressions. For these traits, several genome-wide significant association signals were mapped. To infer the suitability of F2 data to map genes in a segregating breeding population, GWAS results of a pooled F2 cross were validated in two samples of the German Piétrain population (chapter 4). All individuals were genotyped using standard 62k SNP chips. The pooled cross contained the data of two F2 crosses, both had Piétrain as one founder breed, and consisted of 595 individuals. Initially, GWASs were conducted in the pooled F2 cross for the production traits dressing yield, carcass length, daily gain and drip loss. Subsequently, QTL core regions around significant trait associated peaks were defined. Finally, SNPs within these core regions were tested for association in the two samples of the current Piétrain population (771 progeny tested boars and 210 sows) in order to validate them in this breed. In total, 15 QTL were mapped and 8 (5) of them were validated in the boar (sow) validation dataset. This approach takes advantage of the high mapping power in F2 data to detect QTL that may not be found in the segregating Piétrain population. The findings showed that many of the QTL mapped in F2 crosses derived from Piétrain still segregate in this breed, and thus, these F2 datasets provide a promising database to map QTL in the Piétrain breed. The thesis ends with a general discussion.