Insights into the Genetic Architecture Underlying Plasma Lipids and Related Phenotypes from Genome-wide Human Genetic Variation.

Abstract

Complex traits are multifactorial, often with risk contributions from numerous common and rare genetic mutations. The considerable challenges in understanding complex human phenotypes have prompted genome-wide association studies (GWAS), which generally compare large samples of unrelated individuals to test the relationship between genetic markers or nearby linked alleles and modulation of a trait or disease risk. Heritable levels of plasma lipids can influence heart disease risk, highlighting lipid-associated genetic variants as effective therapeutic targets. In collaboration with the Global Lipids Genetics Consortium, I present a follow-up study of approximately 100,000 individuals genotyped on a custom Metabochip array in the largest meta-analysis for lipids to-date. I report 62 novel genetic loci associated with lipids and present downstream bioinformatics analyses to support the role of these loci in lipid regulation. Many of the GWAS-identified lipid loci are non-protein-coding, suggesting a role in transcriptional regulation. This regulatory role can involve altering the DNA sequence at which proteins bind, ultimately affecting gene expression levels in particular cell types. I developed an open source tool called GREGOR (Genomic Regulatory Elements and Gwas Overlap AlgoRithm) to evaluate enrichment of GWAS variants in tissue-specific regulatory features defined by experimental approaches such as chromatin immunoprecipitation followed by high-throughput DNA sequencing (ChIP-seq). I report strong evidence for enrichment in DNase hypersensitive sites of biologically relevant tissues for 5 phenotypes including lipids, coronary artery disease, blood pressure, body mass index, and type 2 diabetes. In addition, I evaluate regulatory feature overlap of linked variants at a set of individual lipid-associated loci to predict the functionality of particular variants, and present experimental results to support my computational predictions. Lastly, I perform discovery and genotyping of structural variation (SV) from low-pass whole genome sequence data of 2,202 Norwegian cases with early-onset myocardial infarction (MI) and matched controls. I use complementary and established SV detection algorithms to call deletions, duplications, and inversions, and perform association analyses with MI disease risk and lipid levels. I observe a deletion in strong linkage disequilibrium with a known MI-associated single variant at the WDR12 locus, suggesting its plausibility as a functional variant at that locus.