Characteristics of DNA methylation and gene expression in regulatory features on the Infinium 450k Beadchip

Abstract

Understanding the relationship between variations in DNA methylation and gene expression has been challenging. Evidence suggests the function of DNA methylation may vary with genomic context, and few consistent rules linking methylation to expression have been noted. For array-based studies, the content of current DNA methylation array platforms provide broad coverage of the genome but target only a fraction of the potentially methylated CG dinucleotides. A better understanding of the interplay between DNA methylation and gene expression is beneficial for users of these platforms, and may aid with candidate prioritization in epigenome-wide association studies (EWAS). To address this we examined the relationship between DNA methylation levels and gene expression in primary T-lymphocytes at discreet genomic regions around the transcriptional unit (Promoters, gene body, untranslated regions) and at CpG island-associated regions (islands, shores and shelves), stratifying by high and low expressed genes. As anticipated we found evidence that DNA methylation at CpG sites near promoter regions are tightly correlated with gene expression in both the stably expressed and developmentally regulated genes, however this is dependent on CpG density. DNA methylation within the gene body was not consistently associated with changes in gene expression. CpG islands and island shores exhibited strong correlations with gene expression, but this was not true for island shelves. We found these relationships were generally preserved at both dynamic and steady state genes, with some notable exceptions. In combination these insights may be useful for prioritising candidates identified in epigenome-wide association studies for subsequent functional studies.