Abstract
Nearly all trait-associated variants identified in GWAS are non-coding. The cis regulatory effects of these variants have been extensively characterized, but how they impact gene regulation in trans has been the subject of much fewer studies. Mapping trans genetic effects is very challenging because their effect sizes tend to be small and a large multiple testing burden reduces the power to detect them. In addition, read mapping biases can lead to many false positives. To reduce mapping biases and substantially improve power to map trans-eQTLs, we developed a pipeline called trans-PCO, which combines careful read and gene filters with a principal component (PC)-based multivariate association test. Our simulations demonstrate that trans-PCO substantially outperforms existing trans-eQTL mapping methods, including univariate and primary PC-based methods. We applied trans-PCO to two gene expression datasets from whole blood, DGN (N = 913) and eQTLGen (N = 31,684), to identify trans-eQTLs associated with gene co-expression networks and hallmark gene sets representing well-defined biological processes. In total, we identified 14,985 high-quality trans-eQTLs associated with 197 co-expression gene modules and biological processes. To better understand the effects of trait-associated variants on gene regulatory networks, we performed colocalization analyses between GWAS loci of 46 complex traits and trans-eQTLs identified in DGN. We highlight several examples where our map of trans effects helps us understand how trait-associated variants impact gene regulatory networks and biological pathways. For example, we found that a locus associated with platelet traits near ARHGEF3 trans-regulates a set of co-expressed genes significantly enriched in the platelet activation pathway. Additionally, six red blood cell trait-associated loci trans-regulate a gene set representing heme metabolism, a crucial process in erythropoiesis. In conclusion, trans-PCO is a powerful and reliable tool that detects trans regulators of cellular pathways and networks, which opens up new opportunities to learn the impact of trait-associated loci on gene regulatory networks.
Competing Interest Statement
The authors have declared no competing interest.
