Utilising multi-omic epidemiology approaches to better understand the molecular pathways underpinning glioma onset.

Abstract

Mendelian randomization (MR) is a robust causal inference technique that utilises molecular data to assess the association between an exposure and an outcome of interest. This method can therefore enhance the understanding of risk factors and underlying molecular pathways that are relevant in the aetiology of a disease. To date, few studies have utilised this approach to evaluate the association between risk factors and glioma risk.
In this thesis, I have applied MR along with other epidemiological techniques (text mining, polygenic risk scores and colocalization) to appraise the role of risk factors with observational evidence (Chapter 3), DNA methylation (Chapter 4), tissue-specific gene expression (Chapter 5), single-cell gene expression and single-cell alternative splicing (Chapter 6), in glioma risk.
Analyses undertaken in Chapter 3 provided evidence for a role of genetically-proxied increased telomere length in the development of glioma. Chapter 4 presented evidence of novel
associations between DNA methylation variation at three cytosine-guanine dinucleotides (CpG) sites with glioma risk. However, inferences were limited by potential tissue-specific effects as the genetic proxies for DNA methylation were derived from blood tissue. In Chapter 5, evidence was provided to support the causal role of differential expression of seven genes in blood tissue in the aetiology of glioma. In addition, novel differential expression of two genes in brain tissue was associated with glioma risk. These findings highlighted that gene expression-glioma associations differed depending on the tissue examined. Finally, in Chapter 6 a novel association between oligodendrocyte-specific CEP70 expression with glioma risk was identified using brain cell type-specific gene expression data. There was little evidence for the role of single-cell alternative splicing in glioma risk. These findings indicate that a deeper understanding of regulatory pathways underpinning glioma risk can be gained by refining the cellular context in which omic data are derived.

Date of Award	9 May 2023
Original language	English
Awarding Institution	University of Bristol
Supervisor	Kathreena M Kurian (Supervisor), Jie Zheng (Supervisor), Caroline Relton (Supervisor), Richard M Martin (Supervisor), James Yarmolinsky (Supervisor) & Philip C Haycock (Supervisor)