Leveraging genome-wide data to understand cardiovascular diseases and their risk factors

Abstract

Cardiovascular diseases (CVD) refer to a group of diseases, such as coronary heart disease, that affect the heart and blood vessels. They contribute significantly to death and illness, resulting in financial costs for society and impacting quality of life. The risk factors for CVDs are multifactorial, and the key modifiable risk factors include elevated cholesterol and blood pressure, diabetes, smoking, diet, physical activity, and obesity. This thesis utilized recent genome-wide association data and methodological advancements to deepen the understanding of genetic factors impacting various non-communicable diseases, specifically focusing on cardiovascular diseases and their related risk factors. The shared genetic architecture between periodontal diseases (PER) and glycemic traits were explored, including Type 2 diabetes (T2D), using genetic correlations, cross-phenotype association, expression-trait association, and functional analysis, such as pathway analysis. The findings revealed a genetic link between them, with a significant positive global genetic correlation. Through local genetic correlation analysis, several genomic regions were identified as shared between PER and glycemic traits or T2D. Additionally, 62 independent pleiotropic loci impacting both PER and glycemic traits were identified, including T2D. The genetically predicted liability of Homeostasis Model Assessment of Beta-cell function adjusted for body mass index (BMI) was found to be causally associated with the risk of PER. The shared genetic etiology between PER and blood lipid traits were investigated using similar approaches as Chapter 3. The study provided insights into the shared basis and novel biological mechanisms between them. A significant global genetic correlation was found between PER and blood lipid traits. Moreover, 58 independent pleiotropic loci were shared between PER and blood lipid traits. Two shared expression-trait associations were observed for total cholesterol or triglyceride and PER in coronary artery and cerebellar hemisphere tissues. The causal relationship between two cardiovascular proteins, Galectin-3 (gal-3) and Suppression of Tumorigenicity 2 (ST2) Protein, and sleep traits and cardiometabolic traits were explored. The findings showed that genetically predicted gal-3 was significantly or suggestively causally associated with certain sleep traits. Reverse mendelian randomization (MR) analysis revealed significant or suggestive associations between genetically predicted gal-3 and triglyceride, fasting insulin, glycated hemoglobin (HbA1c) and Waist-to-Hip Ratio adjusted for BMI. Additionally, genetically predicted ST2 was found to be suggestively associated with sleep duration. In the reverse analysis, triglyceride and sleep duration were suggestively associated with genetically predicted ST2. The causal association between amyotrophic lateral sclerosis (ALS) and 14 cardiovascular conditions were inspected using MR approaches. The results suggested a causal association between ALS and certain cardiovascular conditions and significant or suggestive associations between certain cardiovascular conditions and ALS. Multivariable MR analysis demonstrated that myocardial infarction was significantly associated with the risk of ALS after controlling for low-density lipoprotein cholesterol. In summary, my thesis applied recent genetic association statistical methods to investigate the genetic overlap between cardiovascular diseases or their risk factors and other complex traits. It provides a better insight into the biological mechanisms and genetic architecture underlying cardiovascular diseases and their risk factors. Further biological experiments are expected to consolidate these statistically and computationally supported results. (Word count: 498 words)