Research Questions and Objectives:
Cardiovascular-cerebrovascular disease (CVD/CerVD) is a leading cause of mortality and disability worldwide. While traditional risk factors (e.g., hypertension, diabetes, smoking) are well-established, the molecular and structural underpinnings of disease onset and progression remain incompletely understood. Leveraging UK Biobank’s comprehensive multi-omics data-genomics, proteomics, metabolomics, and deep imaging-we aim to: (1) Identify integrated biomarker signatures (genetic, proteomic, metabolic, imaging) associated with CVD/CerVD and its key risk factors; (2) Elucidate biological pathways linking metabolic dysregulation, chronic inflammation, endothelial dysfunction, and subclinical organ damage; (3) Develop AI-powered, multi-modal risk models for early detection and personalized prevention.
Scientific Rationale:
CVD/CerVD arises from complex interactions between genetic susceptibility, metabolic disturbances, and environmental exposures. UK Biobank offers unparalleled resources: genome-wide SNPs, plasma proteomics and metabolomics, brain and cardiac MRI, carotid intima-media thickness, and detailed phenotypic data. We will construct polygenic risk scores (PRS) and integrate them with circulating biomarkers (e.g., inflammatory proteins, lipid species) and imaging-derived phenotypes (e.g., white-matter hyperintensities, left ventricular mass, visceral adiposity). Two-sample Mendelian randomization will assess causal relationships between omics traits and CVD/CerVD outcomes. Machine learning models will fuse genomic, proteomic, metabolomic, and imaging data to improve risk prediction beyond conventional models. Model performance will be evaluated using cross-validation and sensitivity analyses. This integrative, systems-level approach will uncover novel biological insights into CVD/CerVD pathogenesis and support the development of precision prevention strategies grounded in multi-omics evidence.
