Allele-specific DNA methylation is increased in cancers and its dense mapping in normal plus neoplastic cells increases the yield of disease-associated regulatory SNPs

Catherine Do,Karen Marder,Péter L Nagy ,Govind Bhagat,Angelica Castano,Catherine Monk,Kar Fai Chow ,Lawrence S Honig ,Peter H.r Green ,David S Siegel ,Nicholas P Illsley,Subha Madhavan,Arunjot Singh,Leonel Maldonado,Emmanuel Dumont ,Soren Lehman,Angela M Christiano,Rena Feinman,Sonia Dasilva-Arnold ,George J Kaptain ,André Goy ,Huthayfa Mujahed,Martha Salas,Cornelia L Trimble ,Samuel Goldlust,Benjamin Tycko

doi:10.1186/s13059-020-02059-3

Abstract

BackgroundMapping of allele-specific DNA methylation (ASM) can be a post-GWAS strategy for localizing regulatory sequence polymorphisms (rSNPs). The advantages of this approach, and the mechanisms underlying ASM in normal and neoplastic cells, remain to be clarified.ResultsWe perform whole genome methyl-seq on diverse normal cells and tissues and three cancer types. After excluding imprinting, the data pinpoint 15,112 high-confidence ASM differentially methylated regions, of which 1838 contain SNPs in strong linkage disequilibrium or coinciding with GWAS peaks. ASM frequencies are increased in cancers versus matched normal tissues, due to widespread allele-specific hypomethylation and focal allele-specific hypermethylation in poised chromatin. Cancer cells show increased allele switching at ASM loci, but disruptive SNPs in specific classes of CTCF and transcription factor binding motifs are similarly correlated with ASM in cancer and non-cancer. Rare somatic mutations affecting these same motif classes track with de novo ASM. Allele-specific transcription factor binding from ChIP-seq is enriched among ASM loci, but most ASM differentially methylated regions lack such annotations, and some are found in otherwise uninformative “chromatin deserts.”ConclusionsASM is increased in cancers but occurs by a shared mechanism involving disruptive SNPs in CTCF and transcription factor binding sites in both normal and neoplastic cells. Dense ASM mapping in normal plus cancer samples reveals candidate rSNPs that are difficult to find by other approaches. Together with GWAS data, these rSNPs can nominate specific transcriptional pathways in susceptibility to autoimmune, cardiometabolic, neuropsychiatric, and neoplastic diseases.

Highlights

Mapping of allele-specific DNA methylation (ASM) can be a post-Genome-wide association studies (GWAS) strategy for localizing regulatory sequence polymorphisms
Mapping of high-confidence ASM regions in normal and neoplastic human samples The biological samples in this study are listed in Additional file 1: Table S1, and our approaches for identifying ASM DMRs, testing mechanisms, and nominating diseaseassociated regulatory sequence polymorphisms (rSNPs) are diagrammed in Additional file 2: Figure S1A, B
ASM is increased in cancers due to widespread allele-specific CpG hypomethylation and focal allele-specific CpG hypermethylation in regions of poised chromatin As shown in Fig. 1a, when the numbers of ASM index single nucleotide polymorphisms (SNPs) per sample based on whole genome bisulfite sequencing (WGBS) were normalized to the numbers of informative SNPs and the samples classified by normal vs cancer status, the number of ASM DMRs in the cancers overall was on average 5-fold greater than in the overall group of non-neoplastic samples (Wilcoxon p = 1.7 × 10−7)

Summary

Introduction

Mapping of allele-specific DNA methylation (ASM) can be a post-GWAS strategy for localizing regulatory sequence polymorphisms (rSNPs). The advantages of this approach, and the mechanisms underlying ASM in normal and neoplastic cells, remain to be clarified. Genome-wide association studies (GWAS) have implicated numerous DNA sequence variants, mostly single nucleotide polymorphisms (SNPs) in non-coding regions, as candidates for mediating inter-individual differences in disease susceptibility. To promote GWAS statistical signals to biological true-positives, and to identify the functional sequence variants that underlie these signals, several obstacles need to be overcome. ASM mapping and related post-GWAS approaches such as allele-specific chromatin immunoprecipitation-sequencing (ChIP-seq) [13, 14] can facilitate genome-wide screening for disease-linked rSNPs, which can be prioritized for functional studies. The unique advantages of ASM mapping, and its potential non-redundancy with other post-GWAS mapping methods, remain to be clarified

Methods

Results

Discussion

Conclusion