Widespread effects of DNA methylation and intra-motif dependencies revealed by novel transcription factor binding models

Abstract

Several studies suggested that transcription factor (TF) binding to DNA may be impaired or enhanced by DNA methylation. We present MeDeMo, a toolbox for TF motif analysis that combines information about DNA methylation with models capturing intra-motif dependencies. In a large-scale study using ChIP-seq data for 335 TFs, we identify novel TFs that are affected by DNA methylation. Overall, we find that CpG methylation decreases the likelihood of binding for the majority of TFs. For a considerable subset of TFs, we show that intra-motif dependencies are pivotal for accurately modelling the impact of DNA methylation on TF binding.

Single-cell multi-omics of human clonal hematopoiesis reveals that DNMT3A R882 mutations perturb early progenitor states through selective hypomethylation

“…Indeed, the expression of MYC/MAX target genes was negatively correlated with mean methylation of their binding sites (P = 3.2 x 10 -18 , generalized linear model, Fig. 5i), consistent with prior studies indicating that hypomethylation of binding motifs enhances MYC binding 121,122,139,140 . Thus, our single-cell multi-omics profiling provides a potential model for the observed transcriptional aberration in human DNMT3A mutated CH, supporting enhanced fitness of DNMT3A mutated cells via selective hypomethylation of key hematopoietic TF binding motifs.…”

“…5b). These included key regulators of hematopoiesis such as MYC/MAX, whose activities are known to be negatively impacted by DNA methylation of their binding motifs 121,122 , and were found to have increased target expression in mutated cells (Fig. 3c,d).…”

A base-resolution panorama of the in vivo impact of cytosine methylation on transcription factor binding

“…Our results match what has been reported in these studies [2,4,9,10,16], but also reveals a substantial number of novel TFs that are affected by CpG methylation. Of particular interest, our study revealed a significant number of methyl-plus TFs consistent with in vitro studies [5], in stark contrast to previous methods that have attempted to infer the effect of DNA methylation on TF binding [12][13][14][15]. Notably, a large proportion of methyl-plus TFs belonged to the C2H2-ZF family.…”

“…These include efforts to better distinguish bound from unbound sequences using TF binding models that incorporate CpG methylation status [12, 13], as well as tools that expand the ATGC alphabet by adding symbols for methylated cytosines in order to perform methylation-aware de novo motif discovery [14, 15]. These methods, however, only report whether methylation improves TF binding prediction without delineating the direction of the effect [13], lack the resolution to investigate the effect of methylation of individual intra-motif cytosines [13], and/or do not consider the confounding effects of DNA accessibility and regional methylation level [12–15]. As a result, even some of the most classic methyl-binding TFs, such as CEBPB [2] and KAISO [16], are not detected by these methods [12].…”

“…These methods, however, only report whether methylation improves TF binding prediction without delineating the direction of the effect [13], lack the resolution to investigate the effect of methylation of individual intra-motif cytosines [13], and/or do not consider the confounding effects of DNA accessibility and regional methylation level [12–15]. As a result, even some of the most classic methyl-binding TFs, such as CEBPB [2] and KAISO [16], are not detected by these methods [12].…”