A framework for studying multi-omic risk factors and their interplay: application to coronary artery disease

Transcriptome-wide and proteome-wide association studies (TWAS and PWAS) have identified risk genes across complex diseases. However, the contributions of proximal risk variants and cross-omic interplay remain to be understood.

We propose an integrative framework to characterize disease-associated transcripts and proteins, and apply it to coronary artery disease (CAD). We employed S-PrediXcan on a large-scale genome-wide association study (GWAS) for CAD, with prediction models from GTEx v8 whole blood tissue and the Atherosclerosis Risk in Communities (ARIC) plasma protein. Conditional analyses adjusting for nearby CAD risk variants were performed to retain significant associations. Genes identified by both TWAS and PWAS were subsequently examined for associations with CAD risk factors, and colocalization analyses were performed for expression quantitative trait loci (eQTLs) and protein QTLs (pQTLs).

TWAS identified 294 genes, and PWAS identified 79 genes, with 10 genes common between two analyses: CHMP2B, CLIC4, IL6R, MIF, MXRA7, NME2, NUDT5, PCSK9, TAGLN2 , and WARS . The predicted transcripts and proteins of these genes exhibited consistent associations with CAD risk factors, and colocalization tests revealed shared signals between eQTLs and pQTLs.

Our summary-level data framework enables the construction of multi-omic risk profiles for CAD, advancing the understanding of its genetic etiology through integrated transcriptomic and proteomic analyses.