Fine Mapping of Genetic Variants Influencing Complex Traits in Human

Abstract

Genome-wide association studies (GWAS) have successfully identified thousands of single nucleotide polymorphisms (SNPs) associated with complex traits and diseases of human. Meta-analysis of multiple GWASs has become a popular method since it increases the statistical power and reduces false positive findings. In the genetic field, fine mapping is a process of identifying independently associated variants within a genomic region, which improves our understanding of the causal mechanisms underlying human diseases. However, data sharing among different studies is usually unavailable for meta-analysis of a large number of studies. Fine mapping of genetic variation under the current meta-analysis system is administratively onerous and time consuming. Therefore, statistical approaches which can process fine mapping analysis using meta-analysis summary statistics are assuming increasing importance.
This thesis is concerned with the development of two fine mapping methods using meta-analysis summary statistics: Sequential Sentinel SNP Regional Association Plots (SSS-RAP) and haplotype-based regional association analysis program (HAPRAP). SSS-RAP detects SNPs with independent effects conditional on the top associated signal using meta-analysis summary statistics and summary pair-wise SNPs haplotype frequencies obtained from reference genotype panel. I demonstrate that SSS-RAP is as powerful as conditional analysis and ten model selection methods in individual-level. I applied SSS-RAP to meta-analysis of Electrocardiography (ECG) traits, gallbladder disease (GBD) traits and GIANT BMI database. In addition, HAPRAP is an empirical EM style approach, which extends multiple regression and conditional analysis to meta-analysis levels. I demonstrate that HAPRAP is statistically outperforming existing methods in meta-analysis level. I applied HAPRAP to meta-analysis for ECG traits. Finally, I discuss the position of SSS-RAP and HAPRAP in genetic fine mapping and future direction of genetic fine mapping.

Date of Award	30 Sept 2015
Original language	English
Awarding Institution	University of Bristol
Supervisor	Ian N M Day (Supervisor), Tom R Gaunt (Supervisor) & Santiago Rodriguez (Supervisor)