Abstract TU182: Identification of Epigenomic and Transcriptomic Markers of Coronary Artery Calcium: The Coronary Artery Risk Development in Young Adults (CARDIA) Study

Abstract

Background: Coronary artery calcium (CAC) is a strong predictor of incident atherosclerotic cardiovascular disease (ASCVD) and is routinely used to clarify ASCVD risk. While different omic data types have been independently studied in relation to CAC, integrative multi-omic analyses remain limited, particularly in population-based cohorts. By combining epigenetic and transcriptomic data, we aim to delineate novel pathways that contribute to CAC development. Methods: We included 2,421 CARDIA participants at exam year 25 (mean age 50) with blood leukocyte DNA methylation (DNAm) and bulk RNA-seq data, and contemporaneous computed tomography CAC measurements. CAC outcomes were assessed as: binary (present vs absent), categorical (0, <100, ≥100), and continuous (log-transformed Agatston score). Epigenome-wide association studies (EWAS) were performed to assess associations between DNAm sites and CAC outcomes. Differential gene expression (DGE) analyses were performed to assess associations between transcriptomic markers and CAC outcomes. All models were adjusted for age, sex, race, and CARDIA center. Significant DNAm sites and transcripts shared across all three CAC outcomes were mapped to their respective (most proximal) genes, generating two distinct gene sets—one for epigenomic and one for transcriptomic data. We then compared these gene sets to identify overlapping genes associated with CAC across both omic types. Separately, pathway enrichments were assessed for each EWAS and DGE analysis and compared. Results: We identified 815, 1855, and 1546 DNAm sites and 256, 459, and 505 transcripts significantly associated with binary, categorical, and continuous CAC, respectively (FDR q<0.05). Across the three CAC outcomes, 655 DNAm sites and 198 transcripts were shared. Gene mapping of these shared DNAm sites and transcripts yielded two omic-specific gene sets; comparison of the sets revealed nine overlapping genes (Table 1). Pathway enrichment analyses of aggregated EWAS (Figure 1A) and DGE (Figure 1B) results highlighted prominent biological processes implicated in atherosclerosis. Conclusions: Epigenomic-transcriptomic integration in a population-based cohort identified convergent molecular signatures and pathways associated with CAC, which may represent coordinated epigenetic and transcriptional regulation of CAC. Future multi-omic studies leveraging longitudinal profiling may further clarify molecular mechanisms and inform prevention strategies.