Illuminating The Genetic Basis Of Complex Liver Traits In Humans Via Computational Genomics
This dissertation uses statistical and computational methods in human genetics to further our understanding of the genetic basis of liver traits. Genome-wide association studies (GWAS) have provided researchers with many genomic regions associated with liver phenotypes. However, GWA studies do not directly identify causal risk factors, variants, and genes – therefore, in this dissertation, I use complementary computational and statistical approaches to help identify genes, variants, molecular processes, and risk factors for liver traits with cardiometabolic implications. In the first chapter of this dissertation, I examine the role of genetically-driven differences in alternative splicing in blood lipid level variation by mapping splicing quantitative trait loci (sQTL) and integrating these data with lipid GWAS through colocalization analysis. We find that sQTLs provide information as to the causal variants and genes driving variation and provide a level of granularity that cannot be captured by total gene expression measurements. In the second chapter of this dissertation, I use GWAS data in the recently developed framework of Mendelian Randomization (MR) to better understand the causal risk factors for non-alcoholic fatty liver disease (NAFLD) a complex disease of increasing prevalence and limited treatment options. We find that body mass index and central adiposity have independent effects on NAFLD risk, as does birthweight. We are also the first to show that both causal relationships for body mass index and central adiposity replicate in African American and Hispanic populations, which in turn suggests that the underlying genetics of NAFLD are similar across ancestry groups. In sum, this dissertation improves our understanding of complex liver phenotypes by identifying underlying molecular mechanisms and genes (Chapter 1) and genetically determined risk factors (Chapter 2). Importantly, the results of this report provide researchers and clinicians with new targets for pharmacological and behavioral interventions.