Blood cis-eQTL analysis fails to identify novel association signals among sub-threshold candidates from genome-wide association studies in restless legs syndrome.

Eva C Schulte,Alexander Zimprich,Ingo Fietze,Peter Lichtner,Michael Roden,Cornelius G Bachmann,Claudia Schurmann,Annette Peters,Christian Gieger,Walter Paulus,Holger Prokisch,Claudia Trenkwalder,Matthias Nauck,Thomas Meitinger,Juliane Winkelmann,Christian Herder,Birgit Frauscher,Klaus Berger,Thomas Illig,Birgit Högl,Katharina Schramm,Ulf Schminke,Henry Völzke,Wolfgang H Oertel ,Nadine Groß ,Bertram Müller‐Myhsok ,Karin Stiasny‐Kolster

doi:10.1371/journal.pone.0098092

Abstract

Restless legs syndrome (RLS) is a common neurologic disorder characterized by nightly dysesthesias affecting the legs primarily during periods of rest and relieved by movement. RLS is a complex genetic disease and susceptibility factors in six genomic regions have been identified by means of genome-wide association studies (GWAS). For some complex genetic traits, expression quantitative trait loci (eQTLs) are enriched among trait-associated single nucleotide polymorphisms (SNPs). With the aim of identifying new genetic susceptibility factors for RLS, we assessed the 332 best-associated SNPs from the genome-wide phase of the to date largest RLS GWAS for cis-eQTL effects in peripheral blood from individuals of European descent. In 740 individuals belonging to the KORA general population cohort, 52 cis-eQTLs with pnominal<10−3 were identified, while in 976 individuals belonging to the SHIP-TREND general population study 53 cis-eQTLs with pnominal<10−3 were present. 23 of these cis-eQTLs overlapped between the two cohorts. Subsequently, the twelve of the 23 cis-eQTL SNPs, which were not located at an already published RLS-associated locus, were tested for association in 2449 RLS cases and 1462 controls. The top SNP, located in the DET1 gene, was nominally significant (p<0.05) but did not withstand correction for multiple testing (p = 0.42). Although a similar approach has been used successfully with regard to other complex diseases, we were unable to identify new genetic susceptibility factor for RLS by adding this novel level of functional assessment to RLS GWAS data.

Highlights

Restless legs syndrome (RLS) is a common sensory-motor disorder characterized by dysesthesias affecting the legs, triggered by periods of rest, relieved by movement and occurring mostly during the evening and at night. [1] Consequences are severe sleep disturbances, depression, anxiety and possibly increased cardiovascular risk. [2,3] RLS is a complex polygenic phenotype and genome-wide association studies (GWAS) have identified a total of six genomic loci associated with the disease. [4,5,6,7] Still, the susceptibility alleles known to date only explain about 6.8% of the total heritability [6]
RLS-associated single nucleotide polymorphisms (SNPs) are more likely to be cis-eQTLs To test whether RLS-associated SNPs are more commonly ciseQTLs than those not associated, we compared the number of ciseQTLs among the 332 most significantly associated SNPs from the latest RLS GWAS [6] to the 332 with the worst association p-values
After very stringent linkage disequilibrium (LD) pruning, which was necessary as there was significantly higher LD among the potentially associated SNPs compared to the not associated SNPs, we found evidence for an enrichment of cis-eQTLs in the associated vs. the not-associated SNPs (34 cis-eQTLs among 246 SNPs harboring the most significant association signals vs. 28 cis-eQTLs among 313 SNPs showing the least significant association signals; Fisher’s exact test, one-sided, p,0.05, OR = 1.63)

Summary

Introduction

Restless legs syndrome (RLS) is a common sensory-motor disorder characterized by dysesthesias affecting the legs, triggered by periods of rest, relieved by movement and occurring mostly during the evening and at night. [1] Consequences are severe sleep disturbances, depression, anxiety and possibly increased cardiovascular risk. [2,3] RLS is a complex polygenic phenotype and genome-wide association studies (GWAS) have identified a total of six genomic loci associated with the disease. [4,5,6,7] Still, the susceptibility alleles known to date only explain about 6.8% of the total heritability [6]. [2,3] RLS is a complex polygenic phenotype and genome-wide association studies (GWAS) have identified a total of six genomic loci associated with the disease. It is likely that additional risk loci of weaker effect sizes exist that have not yet been ascertained in the GWAS. It has been shown that single nucleotide polymorphisms (SNPs) associated with complex genetic traits are more likely to have an effect on gene expression and, represent expression quantitative trait loci (eQTLs). We sought to prioritize subthreshold RLS association signals from an RLS GWAS [6] via ciseQTLs in the human blood for follow-up association study seeking to highlight additional genetic factors involved in RLS

Objectives

Methods

Results

Conclusion