CONTRIBUTIONS OF RARE GENETIC VARIANTS TO BIOMARKERS AND DISEASES

Abstract

Most human genetic variants are rare (minor allele frequency, MAF <1%). This thesis investigates the significance of rare coding variants (RV), first with a literature review of vascular dementia (VaD), and subsequently in 31 continuous and 18 binary traits, utilizing whole exome sequence from the UK Biobank (N=167,348 and N=173,688, respectively). This was enabled with the development of the rare variant heritability (RARity) estimator and RARity-β. Genetic determinants of VaD are explored through genome-wide association studies, polygenic risk scores, heritability estimates, and family studies. Complexity and heterogeneity of the disease are highlighted, emphasizing the need for large-scale collaborations and integromics approaches to enhance discoveries. RARity estimates RV heritability (ĥ2RV) without assuming a specific genetic architecture. It revealed a significant loss of heritability (79%) due to gene-level RV aggregation. For 27 traits, ĥ2RV exceeded 5%, with height showing the highest at 21.9%. VaD risk factors such as ApoA-I, BMI, blood pressure, LDL-cholesterol, and triglycerides had ĥ2RV of 4.6% to 9.9%. RARity showed RVs as the source of “missing heritability”, identified 11 new gene-phenotype associations using gene-level heritability estimates, and showed that current pathogenicity predictors do not adequately enrich for RVs contributing to trait variance, indicating a need for better predictive algorithms. RARity-β estimates overall (ĥ2RV-liab) and gene-level heritability of binary traits on a liability scale. Significant ĥ2RV-liab was found for hypothyroidism, asthma, hypercholesterolemia, and essential hypertension, identifying 77 genes with significant contributions to ĥ2RV-liab, including 70 new gene-trait relationships. The PEPB1 gene's role in atrial fibrillation and the TSHR gene's link to hypothyroidism and sciatica are discussed. Results suggest that genes contributing significantly to ĥ2RV-liab have functional consequences. Overall, this thesis provides novel methodologies and insights into the understanding of complex traits and diseases.