Developing Novel High-dimensional Integrative Mediation Analysis Methods and Tools for Cardiovascular Diseases

About This Grant

Cardiovascular disease (CVD) remains the leading cause of death globally, yet the molecular pathways linking established risk factors to disease outcomes are poorly understood. Large-scale datasets, such as those from the NHLBI TOPMed program, offer unprecedented opportunities to explore these pathways. However, the highdimensional nature of the molecular data presents analytical challenges. This project aims to develop novel statistical methods and tools for high-dimensional mediation analysis to advance the understanding of molecular mechanisms underlying CVD. Specifically, we will quantify shared molecular mechanisms mediating multiple cardiovascular traits driven by common pathways (Aim 1), develop robust mediation methods using genetic instruments to identify causal pathways while accounting for latent confounding and leveraging large-scale datasets (Aim 2), and integrate diverse multi-omics data to reveal complementary biological processes driving CVD and identify pathways driven by the joint and individual structure of the multi-omics data (Aim 3). Our approach introduces a variance-based mediation measure capable of capturing both strong and weak effects, complemented by scalable computation and uncertainty quantification. By addressing critical gaps in current high-dimensional mediation analysis methods, this work will provide insights into CVD etiology, enable the discovery of novel biomarkers/pathways, and inform targeted prevention and intervention strategies. Finally, to facilitate broad adoption in the scientific community, we will disseminate scalable and user-friendly computational tools that implement these advanced methods.