Quantitative Trait Loci (QTL) mapping has been widely used to identify genetic loci attributable to the variation observed in complex traits. In recent years, gene expression phenotypes have emerged as a new type of quantitative trait for which QTL can be mapped. Locating sequence variation that has an effect on gene expression (eQTL) is thought to be a promising way to elucidate the genetic architecture of quantitative traits. This thesis explores a number of methodological aspects of eQTL mapping (also known as “genetical genomics”) and considers some practical strategies for applying this approach to livestock populations. One of the exciting prospects of genetical genomics is that the combination of expression studies with fine mapping of functional trait loci can guide the reconstruction of gene networks. The thesis begins with an analysis in which correlations between gene expression and meat quality traits in pigs are investigated in relation to a pork meat quality QTL previously identified. The influence on power due to factors including sample size and records of matched subjects is discussed. An efficient experimental design for two-colour microarrays is then put forward, and it is shown to be an effective use of microarrays for mapping additive eQTL in outbred crosses under simulation. However, designs optimised for detecting both additive and dominance eQTL are found to be less effective. Data collected from livestock populations usually have a pedigreed structure. Many family-based association mapping methods are rather computationally intensive, hence are time-consuming when analysing very large numbers of traits. The application of a novel family-based association method is demonstrated; it is shown to be fast, accurate and flexible for genetical genomics. Furthermore, the results show that multiple testing correction alone is not sufficient to control type I errors in genetical genomics and that careful data filtering is essential. While it is important to limit false positives, it is desirable not to miss many true signals. A multi-trait analysis based on grouping of functionally related genes is devised to detect some of the signals overlooked by a univariate analysis. Using an inbred rat dataset, 13 loci are identified with significant linkage to gene sets of various functions defined by Gene Ontology. Applying this method to livestock species is possible, but the current level of annotations is a limiting factor. Finally, the thesis concludes with some current opinions on the development of genetical genomics and its impact on livestock genetics research.

This thesis explores the genetic characterization of the mechanisms underlying complex traits in chicken through the use and development of bioinformatics tools. The characterization of quantitative trait loci controlling complex traits has proven to be very challenging. This thesis comprises the study of experimental designs, annotation procedures and functional analyses. These represent some of the main ‘bottlenecks’ involved in the integration of QTLs with the biological interpretation of high-throughput technologies. The thesis begins with an investigation of the bioinformatics tools and procedures available for genome research, briefly reviewing microarray technology and commonly applied experimental designs. A targeted experimental design based on the concept of genetical genomics is then presented and applied in order to study a known functional QTL responsible for chicken body weight. This approach contrasts the gene expression levels of two alternative QTL genotypes, hence narrowing the QTL-phenotype gap, and, giving a direct quantification of the link between the genotypes and the genetic responses. Potential candidate genes responsible for the chicken body weight QTL are identified by using the location of the genes, their expression and biological significance. In order to deal with the multiple sources of information and exploit the data effectively, a systematic approach and a relational database were developed to improve the annotation of the probes of the ARK-Genomics G. gallus 13K v4.0 cDNA array utilized on the experiment. To follow up the investigation of the targeted genetical genomics study, a detailed functional analysis is performed on the dataset. The aim is to identify the downstream effects through the identification of functional variation found in pathways, and secondly to achieve a further characterization of potential candidate genes by using comparative genomics and sequence analyses. Finally the investigation of the body weight QTL syntenic regions and their reported QTLs are presented.

The goal of this research is to construct causal gene networks for genetical genomics experiments using expression Quantitative Trait Loci (eQTL) mapping and Structural Equation Modeling (SEM). Unlike Bayesian Networks, this approach is able to construct cyclic networks, while cyclic relationships are expected to be common in gene networks. Reconstruction of gene networks provides important knowledge about the molecular basis of complex human diseases and generally about living systems. In genetical genomics, a segregating population is expression profiled and DNA marker genotyped. An Encompassing Directed Network (EDN) of causal regulatory relationships among genes can be constructed with eQTL mapping and selection of candidate causal regulators. Several eQTL mapping approaches and local structural models were evaluated in their ability to construct an EDN. The edges in an EDN correspond to either direct or indirect causal relationships, and the EDN is likely to contain cycles or feedback loops. We implemented SEM with genetics algorithms to produce sub-models of the EDN containing fewer edges and being well supported by the data. The EDN construction and sparsification methods were tested on a yeast genetical genomics data set, as well as the simulated data. For the simulated networks, the SEM approach has an average detection power of around ninety percent, and an average false discovery rate of around ten percent. / Ph. D.

<p>Pathologies of sexual development are common in humans and reflect the precarious processes of sex determination and sexual differentiation. The gonad forms as a bipotential organ, and recent results from the Capel lab revealed that it is initially balanced between testis and ovarian fates by opposing and antagonistic signaling networks. In XY embryos, this balance is disrupted by the transient expression of the Y-linked gene, Sry, which activates genes that promote the testis pathway and oppose the ovarian pathway. While the roles of a few genes have been defined by mutation, current evidence suggests that the interactions of many genes and signaling pathways are involved in the establishment of sexual fate. For example, most cases of disorders of sexual development (DSDs) are unexplained by mutations in known sex determination genes. In addition, recent microarray studies in the mouse revealed that nearly half the transcriptome is expressed in the gonad at the time of sex determination (Embryonic day 11.5, or E11.5), and as many as 1,500 genes are expressed in a sexually dimorphic pattern at this early stage. Thus the sexual fate decision in the developing gonad likely depends on a complex network of interacting factors that converge on a critical threshold. </p><p>To begin to elucidate the transcription network topology underlying sex determination, we exploited two inbred mouse strains with well-characterized differences in sex reversal. The common inbred strain C57BL/6J (B6) is uniquely sensitive to XY male-to-female sex reversal in response to a number of genetic perturbations, while other strains, including 129S1/SvImJ (129S1) and DBA/2J (D2) are resistant to sex reversal. We hypothesized that these strain differences in gonad phenotype likely result from underlying expression differences in the gonad at the critical timepoint of E11.5. Using microarrays, we identified significant, reproducible differences in the transcriptome of the E11.5 XY gonad between B6 and 129S1 indicating that the reported sensitivity of B6 to sex reversal is consistent with a higher expression of a female-like transcriptome in B6 XY gonads. Surprisingly, a well-characterized master regulator of the testis pathway, Sox9, was found to be upregulated in the sensitive B6 background, which may serve as a compensatory mechanism to counteract the female-leaning transcriptome and activate the testis pathway in wild type B6 XY gonads.</p><p>We extended our expression analysis to a large set of F2 XY gonads from B6 and 129S1 intercrosses. From each pair of gonads, we analyzed the expression of 56 sex-associated genes by nanoliter-scale quantitative RT-PCR (qRT-PCR). The expression levels of most genes were highly variable across the F2 population, yet strong correlations among genes emerged. We employed a First-Order Conditional Independence (FOCI) algorithm to estimate the F2 coexpression network. From this unbiased analysis of XY expression data, we uncovered two subnetworks consisting of primarily male and female genes. Furthermore, we predicted roles for genes of unknown function based on their connectivity and position within the network. </p><p>To identify the genes responsible for these strain expression differences, we genotyped each F2 embryo at 128 single nucleotide polymorphisms (SNPs) located evenly throughout the 19 autosomes and X chromosome. We then employed linkage analysis to detect autosomal regions that control the expression of one or more of the 56 genes in the F2 population. These regions are termed expression quantitative trait loci, or eQTLs. We identified eQTLs for many sex-related genes, including Sry and Sox9, the key regulators of male sex determination. In addition, we identified multiple prominent trans-band eQTLs that controlled the expression of many genes. My work represents the first eQTL analysis of a developing vertebrate organ, the mouse gonad. This systems-level approach revealed the complex transcription architecture underlying sex determination, and provides a mechanistic explanation for sensitivity to sex reversal seen in some individuals.</p> / Dissertation

The copious information generated from transcriptomes gives us an opportunity to learn biological processes as integrated systems; however, due to numerous sources of variation, high dimensions of data structure, various levels of data quality, and different formats of the inputs, dissecting and interpreting such data presents daunting challenges to scientists. The goal of this research is to provide improved and new statistical tools for analyzing transcriptomes data to identify gene expression patterns for classifying samples, to discover regulatory gene networks using natural genetic perturbations, to develop statistical methods for model fitting and comparison of biochemical networks, and eventually to advance our capability to understand the principles of biological processes at the system level. / Ph. D.