Methylated DNA immunoprecipitation followed by high-throughput sequencing (MeDIP-seq) has the potential to identify changes in DNA methylation important in cancer development. In order to understand the role of epigenetic modulation in the development of acute myeloid leukemia (AML) we have applied MeDIP-seq to the DNA of 12 AML patients and 4 normal bone marrows. This analysis revealed leukemia-associated differentially methylated regions that included gene promoters, gene bodies, CpG islands and CpG island shores. Two genes (SPHKAP and DPP6) with significantly methylated promoters were of interest and further analysis of their expression showed them to be repressed in AML. We also demonstrated considerable cytogenetic subtype specificity in the methylomes affecting different genomic features. Significantly distinct patterns of hypomethylation of certain interspersed repeat elements were associated with cytogenetic subtypes. The methylation patterns of members of the SINE family tightly clustered all leukemic patients with an enrichment of Alu repeats with a high CpG density (P<0.0001). We were able to demonstrate significant inverse correlation between intragenic interspersed repeat sequence methylation and gene expression with SINEs showing the strongest inverse correlation (R(2) = 0.7). We conclude that the alterations in DNA methylation that accompany the development of AML affect not only the promoters, but also the non-promoter genomic features, with significant demethylation of certain interspersed repeat DNA elements being associated with AML cytogenetic subtypes. MeDIP-seq data were validated using bisulfite pyrosequencing and the Infinium array.

Monozygotic (MZ) twin pair discordance for childhood-onset Type 1 Diabetes (T1D) is ∼50%, implicating roles for genetic and non-genetic factors in the aetiology of this complex autoimmune disease. Although significant progress has been made in elucidating the genetics of T1D in recent years, the non-genetic component has remained poorly defined. We hypothesized that epigenetic variation could underlie some of the non-genetic component of T1D aetiology and, thus, performed an epigenome-wide association study (EWAS) for this disease. We generated genome-wide DNA methylation profiles of purified CD14+ monocytes (an immune effector cell type relevant to T1D pathogenesis) from 15 T1D-discordant MZ twin pairs. This identified 132 different CpG sites at which the direction of the intra-MZ pair DNA methylation difference significantly correlated with the diabetic state, i.e. T1D-associated methylation variable positions (T1D-MVPs). We confirmed these T1D-MVPs display statistically significant intra-MZ pair DNA methylation differences in the expected direction in an independent set of T1D-discordant MZ pairs (P = 0.035). Then, to establish the temporal origins of the T1D-MVPs, we generated two further genome-wide datasets and established that, when compared with controls, T1D-MVPs are enriched in singletons both before (P = 0.001) and at (P = 0.015) disease diagnosis, and also in singletons positive for diabetes-associated autoantibodies but disease-free even after 12 years follow-up (P = 0.0023). Combined, these results suggest that T1D-MVPs arise very early in the etiological process that leads to overt T1D. Our EWAS of T1D represents an important contribution toward understanding the etiological role of epigenetic variation in type 1 diabetes, and it is also the first systematic analysis of the temporal origins of disease-associated epigenetic variation for any human complex disease.

Pluripotent stem cells (PSCs) occupy a spectrum of reversible molecular states ranging from a naive ground-state in 2i, to metastable embryonic stem cells (ESCs) in serum, to lineage-primed epiblast stem cells (EpiSCs). To investigate the role of DNA methylation (5mC) across distinct pluripotent states, we mapped genome-wide 5mC and 5-hydroxymethycytosine (5hmC) in multiple PSCs. Ground-state ESCs exhibit an altered distribution of 5mC and 5hmC at regulatory elements and dramatically lower absolute levels relative to ESCs in serum. By contrast, EpiSCs exhibit increased promoter 5mC coupled with reduced 5hmC, which contributes to their developmental restriction. Switch to 2i triggers rapid onset of both the ground-state gene expression program and global DNA demethylation. Mechanistically, repression of de novo methylases by PRDM14 drives DNA demethylation at slow kinetics, whereas TET1/TET2-mediated 5hmC conversion enhances both the rate and extent of hypomethylation. These processes thus act synergistically during transition to ground-state pluripotency to promote a robust hypomethylated state.

In neurons, the timely and accurate expression of genes in response to synaptic activity relies on the interplay between epigenetic modifications of histones, recruitment of regulatory proteins to chromatin and changes to nuclear structure. To identify genes and regulatory elements responsive to synaptic activation in vivo, we performed a genome-wide ChIPseq analysis of acetylated histone H3 using somatosensory cortex of mice exposed to novel enriched environmental (NEE) conditions. We discovered that Short Interspersed Elements (SINEs) located distal to promoters of activity-dependent genes became acetylated following exposure to NEE and were bound by the general transcription factor TFIIIC. Importantly, under depolarizing conditions, inducible genes relocated to transcription factories (TFs), and this event was controlled by TFIIIC. Silencing of the TFIIIC subunit Gtf3c5 in non-stimulated neurons induced uncontrolled relocation to TFs and transcription of activity-dependent genes. Remarkably, in cortical neurons, silencing of Gtf3c5 mimicked the effects of chronic depolarization, inducing a dramatic increase of both dendritic length and branching. These findings reveal a novel and essential regulatory function of both SINEs and TFIIIC in mediating gene relocation and transcription. They also suggest that TFIIIC may regulate the rearrangement of nuclear architecture, allowing the coordinated expression of activity-dependent neuronal genes.

Recent multi-dimensional approaches to the study of complex disease have revealed powerful insights into how genetic and epigenetic factors may underlie their aetiopathogenesis. We examined genotype-epigenotype interactions in the context of Type 2 Diabetes (T2D), focussing on known regions of genomic susceptibility. We assayed DNA methylation in 60 females, stratified according to disease susceptibility haplotype using previously identified association loci. CpG methylation was assessed using methylated DNA immunoprecipitation on a targeted array (MeDIP-chip) and absolute methylation values were estimated using a Bayesian algorithm (BATMAN). Absolute methylation levels were quantified across LD blocks, and we identified increased DNA methylation on the FTO obesity susceptibility haplotype, tagged by the rs8050136 risk allele A (p = 9.40×10(-4), permutation p = 1.0×10(-3)). Further analysis across the 46 kb LD block using sliding windows localised the most significant difference to be within a 7.7 kb region (p = 1.13×10(-7)). Sequence level analysis, followed by pyrosequencing validation, revealed that the methylation difference was driven by the co-ordinated phase of CpG-creating SNPs across the risk haplotype. This 7.7 kb region of haplotype-specific methylation (HSM), encapsulates a Highly Conserved Non-Coding Element (HCNE) that has previously been validated as a long-range enhancer, supported by the histone H3K4me1 enhancer signature. This study demonstrates that integration of Genome-Wide Association (GWA) SNP and epigenomic DNA methylation data can identify potential novel genotype-epigenotype interactions within disease-associated loci, thus providing a novel route to aid unravelling common complex diseases.

DNase I is an enzyme which cuts duplex DNA at a rate that depends strongly upon its chromatin environment. In combination with high-throughput sequencing (HTS) technology, it can be used to infer genome-wide landscapes of open chromatin regions. Using this technology, systematic identification of hundreds of thousands of DNase I hypersensitive sites (DHS) per cell type has been possible, and this in turn has helped to precisely delineate genomic regulatory compartments. However, to date there has been relatively little investigation into possible biases affecting this data.

Inter-individual epigenetic variation, due to genetic, environmental or random influences, is observed in many eukaryotic species. In mammals, however, the molecular nature of epiallelic variation has been poorly defined, partly due to the restricted focus on DNA methylation. Here we report the first genome-scale investigation of mammalian epialleles that integrates genomic, methylomic, transcriptomic and histone state information.

NestedMICA is a new, scalable, pattern-discovery system for finding transcription factor binding sites and similar motifs in biological sequences. Like several previous methods, NestedMICA tackles this problem by optimizing a probabilistic mixture model to fit a set of sequences. However, the use of a newly developed inference strategy called Nested Sampling means NestedMICA is able to find optimal solutions without the need for a problematic initialization or seeding step. We investigate the performance of NestedMICA in a range scenario, on synthetic data and a well-characterized set of muscle regulatory regions, and compare it with the popular MEME program. We show that the new method is significantly more sensitive than MEME: in one case, it successfully extracted a target motif from background sequence four times longer than could be handled by the existing program. It also performs robustly on synthetic sequences containing multiple significant motifs. When tested on a real set of regulatory sequences, NestedMICA produced motifs which were good predictors for all five abundant classes of annotated binding sites.

A key step in understanding gene regulation is to identify the repertoire of transcription factor binding motifs (TFBMs) that form the building blocks of promoters and other regulatory elements. Identifying these experimentally is very laborious, and the number of TFBMs discovered remains relatively small, especially when compared with the hundreds of transcription factor genes predicted in metazoan genomes. We have used a recently developed statistical motif discovery approach, NestedMICA, to detect candidate TFBMs from a large set of Drosophila melanogaster promoter regions. Of the 120 motifs inferred in our initial analysis, 25 were statistically significant matches to previously reported motifs, while 87 appeared to be novel. Analysis of sequence conservation and motif positioning suggested that the great majority of these discovered motifs are predictive of functional elements in the genome. Many motifs showed associations with specific patterns of gene expression in the D. melanogaster embryo, and we were able to obtain confident annotation of expression patterns for 25 of our motifs, including eight of the novel motifs. The motifs are available through Tiffin, a new database of DNA sequence motifs. We have discovered many new motifs that are overrepresented in D. melanogaster promoter regions, and offer several independent lines of evidence that these are novel TFBMs. Our motif dictionary provides a solid foundation for further investigation of regulatory elements in Drosophila, and demonstrates techniques that should be applicable in other species. We suggest that further improvements in computational motif discovery should narrow the gap between the set of known motifs and the total number of transcription factors in metazoan genomes.

Mouse Embryonic Stem (ES) cells express a unique set of microRNAs (miRNAs), the miR-290-295 cluster. To elucidate the role of these miRNAs and how they integrate into the ES cell regulatory network requires identification of their direct regulatory targets. The difficulty, however, arises from the limited complementarity of metazoan miRNAs to their targets, with the interaction requiring as few as six nucleotides of the miRNA seed sequence. To identify miR-294 targets, we used Dicer1-null ES cells, which lack all endogenous mature miRNAs, and introduced just miR-294 into these ES cells. We then employed two approaches to discover miR-294 targets in mouse ES cells: transcriptome profiling using microarrays and a biochemical approach to isolate mRNA targets associated with the Argonaute2 (Ago2) protein of the RISC (RNA Induced Silencing Complex) effector, followed by RNA-sequencing. In the absence of Dicer1, the RISC complexes are largely devoid of mature miRNAs and should therefore contain only transfected miR-294 and its base-paired targets. Our data suggest that miR-294 may promote pluripotency by regulating a subset of c-Myc target genes and upregulating pluripotency-associated genes such as Lin28.

DNA methylation can regulate gene expression by modulating the interaction between DNA and proteins or protein complexes. Conserved consensus motifs exist across the human genome ("predicted transcription factor binding sites": "predicted TFBS") but the large majority of these are proven by chromatin immunoprecipitation and high throughput sequencing (ChIP-seq) not to be biological transcription factor binding sites ("empirical TFBS"). We hypothesize that DNA methylation at conserved consensus motifs prevents promiscuous or disorderly transcription factor binding.

Most vertebrate promoters lie in unmethylated CpG-dense islands, whereas methylation of the more sparsely distributed CpGs in the remainder of the genome is thought to contribute to transcriptional repression. Nonmethylated CG dinucleotides are recognized by CXXC finger protein 1 (CXXC1, also known as CFP1), which recruits SETD1A (also known as Set1) methyltransferase for trimethylation of histone H3 lysine 4, an active promoter mark. Genomic regions enriched for CpGs are thought to be either absent or irrelevant in invertebrates that lack DNA methylation, such as C. elegans; however, a CXXC1 ortholog (CFP-1) is present. Here we demonstrate that C. elegans CFP-1 targets promoters with high CpG density, and these promoters are marked by high levels of H3K4me3. Furthermore, as for mammalian promoters, high CpG content is associated with nucleosome depletion irrespective of transcriptional activity. We further show that highly occupied target (HOT) regions identified by the binding of a large number of transcription factors are CpG-rich promoters in C. elegans and human genomes, suggesting that the unusually high factor association at HOT regions may be a consequence of CpG-linked chromatin accessibility. Our results indicate that nonmethylated CpG-dense sequence is a conserved genomic signal that promotes an open chromatin state, targeting by a CXXC1 ortholog, and H3K4me3 modification in both C. elegans and human genomes.

Genome-wide profiling of open chromatin regions using DNase I and high-throughput sequencing (DNase-seq) is an increasingly popular approach for finding and studying regulatory elements. A variety of algorithms have been developed to identify regions of open chromatin from raw sequence-tag data, which has motivated us to assess and compare their performance. In this study, four published, publicly available peak calling algorithms used for DNase-seq data analysis (F-seq, Hotspot, MACS and ZINBA) are assessed at a range of signal thresholds on two published DNase-seq datasets for three cell types. The results were benchmarked against an independent dataset of regulatory regions derived from ENCODE in vivo transcription factor binding data for each particular cell type. The level of overlap between peak regions reported by each algorithm and this ENCODE-derived reference set was used to assess sensitivity and specificity of the algorithms. Our study suggests that F-seq has a slightly higher sensitivity than the next best algorithms. Hotspot and the ChIP-seq oriented method, MACS, both perform competitively when used with their default parameters. However the generic peak finder ZINBA appears to be less sensitive than the other three. We also assess accuracy of each algorithm over a range of signal thresholds. In particular, we show that the accuracy of F-Seq can be considerably improved by using a threshold setting that is different from the default value.

Discovering overrepresented patterns in amino acid sequences is an important step in protein functional element identification. We adapted and extended NestedMICA, an ab initio motif finder originally developed for finding transcription binding site motifs, to find short protein signals, and compared its performance with another popular protein motif finder, MEME. NestedMICA, an open source protein motif discovery tool written in Java, is driven by a Monte Carlo technique called Nested Sampling. It uses multi-class sequence background models to represent different "uninteresting" parts of sequences that do not contain motifs of interest. In order to assess NestedMICA as a protein motif finder, we have tested it on synthetic datasets produced by spiking instances of known motifs into a randomly selected set of protein sequences. NestedMICA was also tested using a biologically-authentic test set, where we evaluated its performance with respect to varying sequence length.

Development of high-throughput methods for measuring DNA interactions of transcription factors together with computational advances in short motif inference algorithms is expanding our understanding of transcription factor binding site motifs. The consequential growth of sequence motif data sets makes it important to systematically group and categorise regulatory motifs. It has been shown that there are familial tendencies in DNA sequence motifs that are predictive of the family of factors that binds them. Further development of methods that detect and describe familial motif trends has the potential to help in measuring the similarity of novel computational motif predictions to previously known data and sensitively detecting regulatory motifs similar to previously known ones from novel sequence.

In addition to known protein-coding genes, large amounts of apparently non-coding sequence are conserved between the human and mouse genomes. It seems reasonable to assume that these conserved regions are more likely to contain functional elements than less-conserved portions of the genome.

A major concern in common disease epigenomics is distinguishing causal from consequential epigenetic variation. One means of addressing this issue is to identify the temporal origins of epigenetic variants via longitudinal analyses. However, prospective birth-cohort studies are expensive and time consuming. Here, we report DNA methylomics of archived Guthrie cards for the retrospective longitudinal analyses of in-utero-derived DNA methylation variation. We first validate two methodologies for generating comprehensive DNA methylomes from Guthrie cards. Then, using an integrated epigenomic/genomic analysis of Guthrie cards and follow-up samplings, we identify interindividual DNA methylation variation that is present both at birth and 3 yr later. These findings suggest that disease-relevant epigenetic variation could be detected at birth, i.e., before overt clinical disease. Guthrie card methylomics offers a potentially powerful and cost-effective strategy for studying the dynamics of interindividual epigenomic variation in a range of common human diseases.

Epigenetic mechanisms such as microRNA and histone modification are crucially responsible for dysregulated gene expression in heart failure. In contrast, the role of DNA methylation, another well-characterized epigenetic mark, is unknown. In order to examine whether human cardiomyopathy of different etiologies are connected by a unifying pattern of DNA methylation pattern, we undertook profiling with ischaemic and idiopathic end-stage cardiomyopathic left ventricular (LV) explants from patients who had undergone cardiac transplantation compared to normal control. We performed a preliminary analysis using methylated-DNA immunoprecipitation-chip (MeDIP-chip), validated differential methylation loci by bisulfite-(BS) PCR and high throughput sequencing, and identified 3 angiogenesis-related genetic loci that were differentially methylated. Using quantitative RT-PCR, we found that the expression of these genes differed significantly between CM hearts and normal control (p<0.01). Moreover, for each individual LV tissue, differential methylation showed a predicted correlation to differential expression of the corresponding gene. Thus, differential DNA methylation exists in human cardiomyopathy. In this series of heterogeneous cardiomyopathic LV explants, differential DNA methylation was found in at least 3 angiogenesis-related genes. While in other systems, changes in DNA methylation at specific genomic loci usually precede changes in the expression of corresponding genes, our current findings in cardiomyopathy merit further investigation to determine whether DNA methylation changes play a causative role in the progression of heart failure.

Genomic sequences obtained through high-throughput sequencing are not uniformly distributed across the genome. For example, sequencing data of total genomic DNA show significant, yet unexpected enrichments on promoters and exons. This systematic bias is a particular problem for techniques such as chromatin immunoprecipitation, where the signal for a target factor is plotted across genomic features. We have focused on data obtained from Illumina's Genome Analyser platform, where at least three factors contribute to sequence bias: GC content, mappability of sequencing reads, and regional biases that might be generated by local structure. We show that relying on input control as a normalizer is not generally appropriate due to sample to sample variation in bias. To correct sequence bias, we present BEADS (bias elimination algorithm for deep sequencing), a simple three-step normalization scheme that successfully unmasks real binding patterns in ChIP-seq data. We suggest that this procedure be done routinely prior to data interpretation and downstream analyses.

RNA polymerase transcription initiation sites are largely unknown in Caenorhabditis elegans. The initial 5' end of most protein-coding transcripts is removed by trans-splicing, and noncoding initiation sites have not been investigated. We characterized the landscape of RNA Pol II transcription initiation, identifying 73,500 distinct clusters of initiation. Bidirectional transcription is frequent, with a peak of transcriptional pairing at 120 bp. We assign transcription initiation sites to 7691 protein-coding genes and find that they display features typical of eukaryotic promoters. Strikingly, the majority of initiation events occur in regions with enhancer-like chromatin signatures. Based on the overlap of transcription initiation clusters with mapped transcription factor binding sites, we define 2361 transcribed intergenic enhancers. Remarkably, productive transcription elongation across these enhancers is predominantly in the same orientation as that of the nearest downstream gene. Directed elongation from an upstream enhancer toward a downstream gene could potentially deliver RNA polymerase II to a proximal promoter, or alternatively might function directly as a distal promoter. Our results provide a new resource to investigate transcription regulation in metazoans.