Obligatory and facilitative allelic variation in the DNA methylome within common disease-associated loci

Christopher G Bell,Jordana Bell,Kirsten Ward,Yudong Xia,Wei Yuan,Pirro G Hysi,Massimo Mangino,Richard J Acton,Jun Wang,Timothy D Spector,Leonie Roos,Fei Gao

doi:10.1038/s41467-017-01586-1

Abstract

Integrating epigenetic data with genome-wide association study (GWAS) results can reveal disease mechanisms. The genome sequence itself also shapes the epigenome, with CpG density and transcription factor binding sites (TFBSs) strongly encoding the DNA methylome. Therefore, genetic polymorphism impacts on the observed epigenome. Furthermore, large genetic variants alter epigenetic signal dosage. Here, we identify DNA methylation variability between GWAS-SNP risk and non-risk haplotypes. In three subsets comprising 3128 MeDIP-seq peripheral-blood DNA methylomes, we find 7173 consistent and functionally enriched Differentially Methylated Regions. 36.8% can be attributed to common non-SNP genetic variants. CpG-SNPs, as well as facilitative TFBS-motifs, are also enriched. Highlighting their functional potential, CpG-SNPs strongly associate with allele-specific DNase-I hypersensitivity sites. Our results demonstrate strong DNA methylation allelic differences driven by obligatory or facilitative genetic effects, with potential direct or regional disease-related repercussions. These allelic variations require disentangling from pure tissue-specific modifications, may influence array studies, and imply underestimated population variability in current reference epigenomes.

Highlights

Integrating epigenetic data with genome-wide association study (GWAS) results can reveal disease mechanisms
Because of GWAS results co-locating in the same linkage disequilibrium (LD) block, this reduced to a total of 2709 blocks, which cover 22.1% of the entire length of the genome
To identify differentially methylated regions (DMRs) between risk and non-risk GWAS haplotypes, the haplotype-specific DNA methylation (HSM) peak analysis assessed the linear relationship between the allelic count of the GWAS SNP and DNA methylation scores in 500 bp overlapping windows across the LD block

Summary

Introduction

Integrating epigenetic data with genome-wide association study (GWAS) results can reveal disease mechanisms. Our results demonstrate strong DNA methylation allelic differences driven by obligatory or facilitative genetic effects, with potential direct or regional diseaserelated repercussions These allelic variations require disentangling from pure tissue-specific modifications, may influence array studies, and imply underestimated population variability in current reference epigenomes. Integration with diseaserelevant and tissue-specific functional indicators or epigenetic marks within these regions, such as DNase I hypersensitivity sites (DHSs)[2], histone modifications[3, 4] and DNA methylation variation[5, 6], can highlight candidate active variants This dissection of GWAS signals enables progress from associated SNP to mechanistic understanding[7, 8]. Allelic variations in genetic dosage effects on this DNA methylation score are allelic signal differences not epigenetic variability

Methods

Results

Conclusion