Pathway and gene-based analysis of genome wide association studies (GWAS)
Author(s)
Eleftherochorinou, Charikleia
Type
Thesis
Abstract
My PhD thesis comprises the development and application of novel strategies to analyse
genome-wide association studies (GWAS) in the context of functional pathways. I propose
pathway and gene-centric methodologies as complementary tools to the conventional singlemarker
analyses to mine further the GWAS hidden information. I developed the cumulative
trend (CT) test statistic that assesses the cumulative genetic variation of single nucleotide
polymorphisms (SNPs) of genes that interact in the same biological pathway and tests the
association between a disease and the pathway as an entity. I applied this methodology to the
genotypic data of the Wellcome Trust Case Control Consortium (WTCCC) study on Crohn’s
disease (CD), type I diabetes (T1D), rheumatoid arthritis (RA), bipolar disorder, hypertension,
type II diabetes, coronary artery disease; I identified highly significant associations between the
autoimmune diseases (CD, T1D, RA) and inflammatory pathways; almost no association was
identified between the same pathways and the non-inflammatory conditions. I extended my
approach to a pathway-based gene stability selection methodology, which selects associated genes
in the context of associated pathways. This methodology can be used to prioritise genes for
follow up studies. I applied it on two GWAS of RA with different ethnic background and typed
on different platforms and I demonstrated replication at the pathway, gene and in-silico functional
levels. I finally extended my approach on family trios designed GWAS. I applied it on two casecontrol
and family trio datasets of Kawasaki disease (KD). I explored the association between the
TGF-β pathway and KD susceptibility. The involvement of this pathway in KD was further
validated at the gene expression and protein levels. My proposed methodologies were tested on
real datasets and provided reproducible results, which indicates rigor and robustness. I would
therefore suggest their application to single or multiple GWAS as a complement to conventional
single-SNP analysis.
genome-wide association studies (GWAS) in the context of functional pathways. I propose
pathway and gene-centric methodologies as complementary tools to the conventional singlemarker
analyses to mine further the GWAS hidden information. I developed the cumulative
trend (CT) test statistic that assesses the cumulative genetic variation of single nucleotide
polymorphisms (SNPs) of genes that interact in the same biological pathway and tests the
association between a disease and the pathway as an entity. I applied this methodology to the
genotypic data of the Wellcome Trust Case Control Consortium (WTCCC) study on Crohn’s
disease (CD), type I diabetes (T1D), rheumatoid arthritis (RA), bipolar disorder, hypertension,
type II diabetes, coronary artery disease; I identified highly significant associations between the
autoimmune diseases (CD, T1D, RA) and inflammatory pathways; almost no association was
identified between the same pathways and the non-inflammatory conditions. I extended my
approach to a pathway-based gene stability selection methodology, which selects associated genes
in the context of associated pathways. This methodology can be used to prioritise genes for
follow up studies. I applied it on two GWAS of RA with different ethnic background and typed
on different platforms and I demonstrated replication at the pathway, gene and in-silico functional
levels. I finally extended my approach on family trios designed GWAS. I applied it on two casecontrol
and family trio datasets of Kawasaki disease (KD). I explored the association between the
TGF-β pathway and KD susceptibility. The involvement of this pathway in KD was further
validated at the gene expression and protein levels. My proposed methodologies were tested on
real datasets and provided reproducible results, which indicates rigor and robustness. I would
therefore suggest their application to single or multiple GWAS as a complement to conventional
single-SNP analysis.
Date Issued
2011
Date Awarded
2012-01
Copyright Statement
Attribution NoDerivatives 4.0 International Licence (CC BY-ND)
Advisor
Levin, Michael
Coin, Lachlan
Herberg, Jethro
Creator
Eleftherochorinou, Charikleia
Publisher Department
Medicine
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)