Skip to Main content Skip to Navigation
Theses

Development of network-based analysis methods with application to the genetic component of asthma

Abstract : Genome-wide association studies (GWAS) of asthma have been successful in identifying novel asthma-associated loci, but the genes at these loci account only for a part of the whole genetic component. One limitation of GWAS is that they rest on single-marker analyses which are underpowered to detect variants with small marginal effects but rather influence jointly disease risk. To complement the single-marker approaches, more sophisticated strategies, which integrate biological knowledge, such as protein-protein interactions (PPI) or gene networks with GWAS outcomes to identify disease-associated gene modules, have become prominent. The objectives of this thesis were to develop network-based analysis methods, and apply them to asthma GWAS data to identify biological processes and prioritize new candidate genes related to asthma.This thesis consists of two main studies. The first study was to extend an existing network-based method (dmGWAS) to identify novel genes associated with asthma. We used two GWAS datasets, each consisting of the results of a meta-analysis of nine childhood-onset asthma GWAS (5,924 and 6,043 subjects, called META1 and META2, respectively). We developed a novel method to compute gene-level p-values from SNP p-values (fastCGP), and proposed a bi-directional module search method to identify asthma-associated gene modules. Application of these methods to the asthma data detected a gene module of 91 genes significantly associated with asthma (p < 1e-5). This module consisted of a core network and five peripheral subnetworks including high-confidence candidates for asthma. Out of the 91 genes, 19 genes were nominally significant in both META1 and META2 datasets. They included 13 genes at 4 loci previously found associated with asthma (2q12, 5q31, 9p24.1, 17q12-q21), and six genes at six novel loci: CRMP1 (4p16.1), ZNF192 (6p22.1), RAET1E (6q24.3), CTSL1 (9p21.33), C12orf43 (12q24.31) and JAK3 (19p13-p12). Functional analysis of the module revealed four functionally related gene clusters involved in innate and adaptive immunity, chemotaxis, cell-adhesion and transcription regulation, which are biologically meaningful processes underlying asthma risk.The second study of this thesis was to develop a novel network-based method, named SigMod, to search disease-associated gene modules. SigMod takes a list of gene p-values and a gene network as input. It identifies a set of genes that are enriched in high association signals and tend to have strong interconnection via the formulation of a binary quadratic optimization problem. We proposed an algorithm based on graph-cut theory to solve the optimization problem exactly and efficiently. SigMod has several advantages compared to existing methods, including the ability to find the module enriched in highest association signals, the capacity to incorporate edge weights in the network, and the robustness to background noise. Also, the emphasis of selecting strongly interconnected genes can lead to the identification of genes with close functional relevance. We applied SigMod to both simulated and real datasets. This new method outperformed existing approaches. When SigMod was applied to childhood-onset asthma data, it successfully identified a module made of 190 functionally related genes that are biologically relevant for asthma.
Document type :
Theses
Complete list of metadatas

Cited literature [349 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-02466418
Contributor : Abes Star :  Contact
Submitted on : Tuesday, February 4, 2020 - 2:27:17 PM
Last modification on : Friday, August 21, 2020 - 6:53:39 AM
Long-term archiving on: : Tuesday, May 5, 2020 - 4:34:31 PM

File

LIU_Yuanlong_2_va_20171113.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-02466418, version 1

Collections

Citation

Yuanlong Liu. Development of network-based analysis methods with application to the genetic component of asthma. Human genetics. Université Sorbonne Paris Cité, 2017. English. ⟨NNT : 2017USPCC329⟩. ⟨tel-02466418⟩

Share

Metrics

Record views

70

Files downloads

251