Collections > Electronic Theses and Dissertations > PROBING THE ARCHITECTURE OF COMPLEX TRAITS: FUNCTIONAL GENOMICS METHODS AND APPLICATIONS
pdf

Genome wide association studies (GWA) have had tremendous success in identifying genetic variants associated with complex traits. However, the majority of associated loci fall outside of protein coding regions, suggesting a role in regulatory function. This has highlighted a critical need for understanding the regulatory architecture of the genome. Recent advances in high-throughput sequencing technology have enabled transcriptional profiling and mapping of epigenetic features across a broad range of cell types and conditions, both in human and model organisms. As a result, an increasingly higher-resolution genome-wide annotation of regulatory elements is now available. Additionally, expression quantitative trait loci (eQTL) studies mapping the genetic basis of gene expression have identified single nucleotide polymorphisms (SNPs) whose allelic variation correlates with gene expression levels. In conjunction with epigenetic annotations, these results have greatly improved interpretability of variants implicated in complex traits. However, a more comprehensive model of epigenetic regulation in disease can only be obtained by directly assaying disease-relevant tissue in affected individuals. Moreover, traditional eQTL methods often perform a prohibitive number of statistical tests, and are underpowered for detecting weaker associations between SNPs and distally-located genes. In the following chapters I present a novel statistical method that reduces eQTL testing burden and improves power to detect genetic variants associated with expression levels of distal genes. Applying this method to data sets in yeast, mouse, and human, I identified thousands of new eQTL and highlighted candidate master regulators, which were consistently enriched across species for metabolic function. Additionally, I present an analysis of the chromatin and transcriptional landscapes in colon tissue from 33 Crohn’s disease and non-IBD individuals. In ten samples, I found evidence of a molecular signature consistent with metaplasia, the prevalence of which was highly over-represented in CD patients. In an analysis of the remaining individuals, I identified thousands of regulatory regions implicated in disease, many of which co-localize with differentially expressed genes, and highlighted several candidate driver transcription factors. Together, these methods and applications provide a richer understanding of genetic and epigenetic variants implicated in complex traits and disease, and provide hypotheses for future follow up studies.