Thesis (M.A.)--York University, 2001. / Typescript. Includes bibliographical references (leaves 257-270). Also available on the Internet. MODE OF ACCESS via web browser by entering the following URL: http://wwwlib.umi.com/cr/yorku/fullcit?pMQ67717.

Pharmacogenomics offers one of the best use cases for widespread clinical implementation of genomic medicine, as variants tend to have moderate allele frequencies, many of the affected medications are relatively common, and the magnitude of effect tends to be clinically meaningful. Using pharmacogenomic-guided warfarin dosing as an example, this dissertation addresses potential barriers and solutions to the clinical implementation of pharmacogenomics. Warfarin is a blood thinner that has a narrow therapeutic index and wide dosing variation, with many known pharmacogenomic markers associated with stable warfarin dose. A number of different methods to reduce disparities in pharmacogenomic-guided warfarin dosing among African Americans were tested. Additionally data from Vanderbiltâs clinical implementation of pharmacogenomic-guided warfarin dosing were analyzed to assess process outcomes (e.g. how did the actual warfarin dose ordered compare to the recommended dose) and patient outcomes (e.g., what kinds of clinical events did the patient experience immediately following warfarin initiation). A summary of policy and technological challenges for clinical implementation of precision medicine, along with potential solutions, are presented.

Age-related macular degeneration (AMD) is a neurodegenerative disease that leads to a loss of central vision and is the leading cause of blindness in the elderly in developed countries. AMD can significantly reduce quality of life and there is no way to cure or prevent the disease. The prevalence of AMD is expected to increase worldwide as lifespans increase. Of the two subtypes of AMD â geographic atrophy and neovascular AMD â only the neovascular form has treatment options, yet treatment does not reverse vision loss and is required for the life of a patient. AMD is known to be highly heritable (45-70%) but previous genetic studies only explain a portion of the heritability. In this work we perform genetic pathway analyses of AMD to localize additional heritability and develop novel methods to test for cumulative epistatic (interaction) effects potentially influencing risk for AMD. We show that genetic variation in complement activation-related genes, excluding previously associated risk variants, contributes to a statistically significant proportion of risk for AMD (9.7%). Additionally, we develop methods to calculate orthogonal genetic relationship matrices to estimate additive, dominant, and epistatic genetic effects. We applied this method to test for interactions between the ARMS2 gene and three AMD-related pathways and were able to conclusively rule out epistatic effects influencing risk for AMD from the specific interactions that were tested.

Screening for mutations within the BRCA1 and BRCA2 genes is a daunting task due to the length of the genes and the absence of mutational hotspots. An additional contributing factor is the genetic diversity within the Black ethnic groups of Africa, including the Sotho/Tswanas of the Free State. As no information was available regarding the prevalence of BRCA mutations within this group, this pilot study was launched in an attempt to determine the genetic component attributable to BRCA mutations in BC development. The selection criteria were not optimal and possible sporadic or single cases were included which according to literature, is not associated with mutations within the two familial BC genes. This resulted in a low percentage of disease-causing mutations being detected. It is proposed that the selection criteria in the future should emphasize the selection of bilateral BC cases with or without a positive family history. This characteristic seems to be more closely associated with familial BC in the Black patients than an early age at onset. The latter could be masking the familial BC cases, as the median age of onset of the disease in Black ethnic groups is 48. Two disease-causing mutations were identified, one within each of the genes. Both mutations were detected with PTT as they are located within exon 11. This indicates that this technique, although based on older technology, is still a valuable screening technique as it is cost-effective and less time consuming than screening the larger exons with for example high resolution melting (HRM). The two mutations are both situated within critical regions of the genes. BRCA1 c.2069_2072delAAAG,p.Lys653SerfsX699 is located within the binding domain of Rad50 whereas BRCA2 c.6455_6455delT,p.Lys2075ArgfsX2078 is located in BRC repeat 8. The presence of both these mutations will result in a truncated protein that would probably not be able to participate in DNA repair in response to DNA damage and cell cycle control (Green and Lin, 2012). This could result in chromosome instability and therefore tumour formation. Both mutations are novel and have not been detected internationally nor in the Black population residing in Gauteng SA. As all mutations detected thus far for the Black SA population seem to be limited to a single family and with no founder mutations found, full screening of both these genes remains the golden standard. Functional studies should be performed for the intronic variant BRCA2 c.517-4C>G (g.32900632C>G, rs81002804) detected in various patients. As this intronic variant is located only four bp from the start of exon 7, it could play a role in creating an alternative splice site. These studies together with the analysis of control individuals from the various Black SA ethnic groups will resolve the question whether this variant has the potential to be disease-causing.

Lung cancer is classified into two main types: small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC). NSCLC has several subtypes, but the two most commonly occurring subtypes are lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC). These classifications are mostly based upon histological and pathological characteristics, but there is increasing evidence of genetic and molecular differences as well. Although previous work has identified differences between subtypes of NSCLC at the somatic level, little work has been done to look at differences in all three subtypes across both the germline and somatic genomes. I hypothesized that a comprehensive detailed comparison of these lung cancer subtypes at both genomes would reveal shared and distinct mechanisms of disease. In this work, I used single nucleotide polymorphisms (SNPs) identified as significant (p < 1 x 10-3) in a genome wide association study (GWAS) to identify regulatory variants associated with each lung cancer subtype. I used these regulatory SNPs to identify sets of regulated genes in the genome for each subtype. I also expanded this germline work across other lung diseases and cancer types to extend the utility of the pipeline. At the somatic level, I identified genes that were differentially expressed in lung tumor versus normal tissue. Additionally, I identified mutational signatures and sets of potential driver genes in the somatic genomes for each subtype. I also identified biological pathways enriched with the germline and somatic gene sets. Finally, I performed a detailed comparison of calling somatic single nucleotide variants (SNVs) from whole exome sequencing (WES) versus transcriptome sequencing (RNA-Seq) in NSCLC. Overall, across all comparisons and genomes, I observed very little overlap between the three lung cancer subtypes. However, I found one gene, CHRNA5, disrupted by regulatory variants in the germline genomes and differentially expressed in the somatic genomes. This gene encodes a nicotinic acetylcholine receptor and may play a role in lung cancer due to its finding across both genomes and all three subtypes.

While genetic association studies have been able to elucidate the importance of genetics in human disease outcomes, these studies are limited by the necessity of collecting specifically tailored cohorts and that they frequently only test a single outcome. This focus on a single disease at a time ignores the interconnected nature of both biological pathways and disease phenotypes. My dissertation uses phenome-wide association scans (PheWAS), a method of testing one predictor for association with many disease outcomes, to expand our knowledge of multiple genetic variants and types of genetic variation. We used BioVU, a biobank linked to de-identified electronic medical records (EMRs), to explored a variety of applications for PheWAS. Each chapter presents a project where PheWAS was implemented as a starting point due to a specific hypothesis, before follow-up analyses based on the PheWAS outcome and out existing knowledge of the gene, protein, or variant were performed. The projects presented here begin with the most straight-forward scenario, directly genotyped single SNPs, and progress to imputed deletions before exploring ways to use PheWAS in multi-dimensional studies. In conclusion, I used PheWAS to uncover novel genotype-phenotype associations, and further explored these associations using other data types in the EMR. While PheWAS can be a useful tool for discovering unexpected disease consequences of genetic predictors, using it successfully requires sufficient knowledge of the genetic variation tested to evaluate the biological relevance of association signals

Epistasis is a phenomenon wherein the effect of a genetic variant on a phenotype is dependent on the genomic context. Better understanding epistastic relationships between variants, often termed interactions, can shed light on novel genomic loci associated with complex disease, which may improve our understanding of the underlying biological mechanisms. Additionally, capturing epistastic effects in models of disease risk may help improve predictions of at-risk populations, or the prediction of a variantâs deleteriousness in precision medicine initiatives. However, the study of epistasis faces unique methodological challenges, and consequently, evidence for regulatory epistasis remains elusive in humans. In this work, I address two major challenges within the field of regulatory epistasis: the development of statistical best practices, and the investigation of epistasis within haplotypes. In Chapter 2, I illustrate that traditional quality control procedures are insufficient to correct for confounding in studies of epistasis, and develop a set of additional guidelines for future studies. Once these were applied, I found little evidence for epistasis between common, unlinked variants influencing gene expression levels. In Chapter 3, I leverage unique properties of admixed populations to investigate epistasis within ancestral haplotypes disrupted by ancestry-specific recombination events. I find several examples of epistasis with plausible biological support, which serve as a proof of principle for the utility of this approach. Overall, these findings indicate that regulatory epistasis likely has small effects, occurs within haplotypes, or occurs between distant genomic regions; we recommend future studies of epistasis focus on these possibilities.

The genomics era has seen a staggering increase in the number of whole genome sequences. This has bolstered studies of human populations, and revealed regions of the genome bearing signatures of selection and other demographic events. However, tying these regions to phenotypic effects in humans is difficult. I addressed this challenge by leveraging densely phenotyped biobank populations from the eMERGE network, a collection of 10 clinical biobanks across the US that connect electronic health records (EHRs) to genotyping data. The eMERGE data enabled me to interrogate the function of human genetic variation on a broad array of phenotypes using the phenome-wide association study (PheWAS) framework. Each chapter describes a project in which I tested hypotheses about the impact of evolutionarily important variants on human health. In the first, I examine the clinical impact of interbreeding between humans and Neanderthals; in the next, I evaluate variants whose allele frequencies have increased drastically since human divergence from chimpanzee; and finally I consider variants affected by GC-biased gene conversion, a recombination-associated mutational process that favors the fixation of G and C alleles. In conclusion, I used large clinical biobanks to uncover novel genotype-phenotype associations that reveal the effects of recent demographic events and evolutionary processes that have shaped the human genome.

No description available.

No description available.