Open AccessStatistical and Functional Studies Identify Epistasis of Cardiovascular Risk Genomic Variants From Genome‐Wide Association Studies
Author(s) -
Li Yabo,
Cho Hyosuk,
Wang Fan,
CanelaXandri Oriol,
Luo Chunyan,
Rawlik Konrad,
Archacki Stephen,
Xu Chengqi,
Tenesa Albert,
Chen Qiuyun,
Wang Qing Kenneth
Publication year - 2020
Publication title -
journal of the american heart association
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.494
H-Index - 85
ISSN - 2047-9980
DOI - 10.1161/jaha.119.014146
Subject(s) - gene knockdown , genome wide association study , gene expression profiling , gene , biology , genetic association , cancer research , epistasis , genetics , genome editing , gene expression , computational biology , medicine , genome , single nucleotide polymorphism , genotype
Background Epistasis describes how gene‐gene interactions affect phenotypes, and could have a profound impact on human diseases such as coronary artery disease ( CAD ). The goal of this study was to identify gene‐gene interactions in CAD using an easily generalizable multi‐stage approach. Methods and Results Our forward genetic approach consists of multiple steps that combine statistical and functional approaches, and analyze information from global gene expression profiling, functional interactions, and genetic interactions to robustly identify gene‐gene interactions. Global gene expression profiling shows that knockdown of ANRIL (DQ485454) at 9p21.3 GWAS (genome‐wide association studies) CAD locus upregulates TMEM 100 and TMEM 106B . Functional studies indicate that the increased monocyte adhesion to endothelial cells and transendothelial migration of monocytes, 2 critical processes in the initiation of CAD , by ANRIL knockdown are reversed by knockdown of TMEM 106B , but not of TMEM 100 . Furthermore, the decreased monocyte adhesion to endothelial cells and transendothelial migration of monocytes induced by ANRIL overexpression was reversed by overexpressing TMEM 106B . TMEM 106B expression was upregulated by >2‐fold in CAD coronary arteries. A significant association was found between variants in TMEM 106B (but not in TMEM 100 ) and CAD ( P =1.9×10 −8 ). Significant gene‐gene interaction was detected between ANRIL variant rs2383207 and TMEM 106B variant rs3807865 ( P =0.009). A similar approach also identifies significant interaction between rs6903956 in ADTRP and rs17465637 in MIA 3 ( P =0.005). Conclusions We demonstrate 2 pairs of epistatic interactions between GWAS loci for CAD and offer important insights into the genetic architecture and molecular mechanisms for the pathogenesis of CAD . Our strategy has broad applicability to the identification of epistasis in other human diseases.