Nucleotide distance influences co-methylation between nearby CpG sites

Reconstructing DNA methylation maps of ancient populations

Studying premortem DNA methylation from ancient DNA (aDNA) provides a proxy for ancient gene activity patterns, and hence valuable information on evolutionary changes in gene regulation. Due to statistical limitations, current methods to reconstruct aDNA methylation maps are constrained to high-coverage shotgun samples, which comprise a small minority of available ancient samples. Most samples are sequenced using in-situ hybridization capture sequencing which targets a predefined set of genomic positions. Here, we develop methods to reconstruct aDNA methylation maps of samples that were not sequenced using high-coverage shotgun sequencing, by way of pooling together individuals to obtain a DNA methylation map that is characteristic of a population. We show that the resulting DNA methylation maps capture meaningful biological information and allow for the detection of differential methylation across populations. We offer guidelines on how to carry out comparative studies involving ancient populations, and how to control the rate of falsely discovered differentially methylated regions. The ability to reconstruct DNA methylation maps of past populations allows for the development of a whole new frontier in paleoepigenetic research, tracing DNA methylation changes throughout human history, using data from thousands of ancient samples.

Discovery of novel DNA methylation biomarker panels for the diagnosis and differentiation between common adenocarcinomas and their liver metastases

Differentiation between adenocarcinomas is sometimes challenging. The promising avenue for discovering new biomarkers lies in bioinformatics using DNA methylation analysis. Utilizing a 2853-sample identification dataset and a 782-sample independent verification dataset, we have identified diagnostic DNA methylation biomarkers that are hypermethylated in cancer and differentiate between breast invasive carcinoma, cholangiocarcinoma, colorectal cancer, hepatocellular carcinoma, lung adenocarcinoma, pancreatic adenocarcinoma and stomach adenocarcinoma. The best panels for cancer type exhibit sensitivity of 77.8-95.9%, a specificity of 92.7-97.5% for tumors, a specificity of 91.5-97.7% for tumors and normal tissues and a diagnostic accuracy of 85.3-96.4%. We have shown that the results can be extended from the primary cancers to their liver metastases, as the best panels diagnose and differentiate between pancreatic adenocarcinoma liver metastases and breast invasive carcinoma liver metastases with a sensitivity and specificity of 83.3-100% and a diagnostic accuracy of 86.8-91.9%. Moreover, the panels could detect hypermethylation of selected regions in the cell-free DNA of patients with liver metastases. At the same time, these were unmethylated in the cell-free DNA of healthy donors, confirming their applicability for liquid biopsies.

A novel approach toward optimal workflow selection for DNA methylation biomarker discovery

DNA methylation is a major epigenetic modification involved in many physiological processes. Normal methylation patterns are disrupted in many diseases and methylation-based biomarkers have shown promise in several contexts. Marker discovery typically involves the analysis of publicly available DNA methylation data from high-throughput assays. Numerous methods for identification of differentially methylated biomarkers have been developed, making the need for best practices guidelines and context-specific analyses workflows exceedingly high. To this end, here we propose TASA, a novel method for simulating methylation array data in various scenarios. We then comprehensively assess different data analysis workflows using real and simulated data and suggest optimal start-to-finish analysis workflows. Our study demonstrates that the choice of analysis pipeline for DNA methylation-based marker discovery is crucial and different across different contexts.

A novel principal component based method for identifying differentially methylated regions in Illumina Infinium MethylationEPIC BeadChip data

Differentially methylated regions (DMRs) are genomic regions with methylation patterns across multiple CpG sites that are associated with a phenotype. In this study, we proposed a Principal Component (PC) based DMR analysis method for use with data generated using the Illumina Infinium MethylationEPIC BeadChip (EPIC) array. We obtained methylation residuals by regressing the M-values of CpGs within a region on covariates, extracted PCs of the residuals, and then combined association information across PCs to obtain regional significance. Simulation-based genome-wide false positive (GFP) rates and true positive rates were estimated under a variety of conditions before determining the final version of our method, which we have named DMRPC. Then, DMRPC and another DMR method, coMethDMR, were used to perform epigenome-wide analyses of several phenotypes known to have multiple associated methylation loci (age, sex, and smoking) in a discovery and a replication cohort. Among regions that were analysed by both methods, DMRPC identified 50% more genome-wide significant age-associated DMRs than coMethDMR. The replication rate for the loci that were identified by only DMRPC was higher than the rate for those that were identified by only coMethDMR (90% for DMRPC vs. 76% for coMethDMR). Furthermore, DMRPC identified replicable associations in regions of moderate between-CpG correlation which are typically not analysed by coMethDMR. For the analyses of sex and smoking, the advantage of DMRPC was less clear. In conclusion, DMRPC is a new powerful DMR discovery tool that retains power in genomic regions with moderate correlation across CpGs.

Liquid Biopsy in Alzheimer's Disease Patients Reveals Epigenetic Changes in the PRLHR Gene

In recent years, new DNA methylation variants have been reported in genes biologically relevant to Alzheimer's disease (AD) in human brain tissue. However, this AD-specific epigenetic information remains brain-locked and unreachable during patients' lifetimes. In a previous methylome performed in the hippocampus of 26 AD patients and 12 controls, we found higher methylation levels in AD patients in the promoter region of PRLHR, a gene involved in energy balance regulation. Our aim was to further characterize PRLHR's role in AD and to evaluate if the liquid biopsy technique would provide life access to this brain information in a non-invasive way. First, we extended the methylation mapping of PRLHR and validated previous methylome results via bisulfite cloning sequencing. Next, we observed a positive correlation between PRLHR methylation levels and AD-related neuropathological changes and a decreased expression of PRLHR in AD hippocampus. Then, we managed to replicate the hippocampal methylation differences in plasma cfDNA from an additional cohort of 35 AD patients and 35 controls. The isolation of cfDNA from the plasma of AD patients may constitute a source of potential epigenetic biomarkers to aid AD clinical management.

Transposon wave remodeled the epigenomic landscape in the rapid evolution of X-Chromosome dosage compensation

Sex chromosome dosage compensation is a model to understand the coordinated evolution of transcription; however, the advanced age of the sex chromosomes in model systems makes it difficult to study how the complex regulatory mechanisms underlying chromosome-wide dosage compensation can evolve. The sex chromosomes of Poecilia picta have undergone recent and rapid divergence, resulting in widespread gene loss on the male Y, coupled with complete X Chromosome dosage compensation, the first case reported in a fish. The recent de novo origin of dosage compensation presents a unique opportunity to understand the genetic and evolutionary basis of coordinated chromosomal gene regulation. By combining a new chromosome-level assembly of P. picta with whole-genome bisulfite sequencing and RNA-seq data, we determine that the YY1 transcription factor (YY1) DNA binding motif is associated with male-specific hypomethylated regions on the X, but not the autosomes. These YY1 motifs are the result of a recent and rapid repetitive element expansion on the P. picta X Chromosome, which is absent in closely related species that lack dosage compensation. Taken together, our results present compelling support that a disruptive wave of repetitive element insertions carrying YY1 motifs resulted in the remodeling of the X Chromosome epigenomic landscape and the rapid de novo origin of a dosage compensation system.

There's more to it: uncovering genomewide DNA methylation heterogeneity

Tweetable abstract Monitoring changes in methylation heterogeneity can be powerful in detecting disease progression early. This editorial highlights the importance of profiling methylation heterogeneity and identifies existing measures and research gaps.

Nucleotide and codon usage biases involved in the evolution of African swine fever virus: A comparative genomics analysis

Since African swine fever virus (ASFV) replication is closely related to its host's machinery, codon usage of viral genome can be subject to selection pressures. A better understanding of codon usage can give new insights into viral evolution. We implemented information entropy and revealed that the nucleotide usage pattern of ASFV is significantly associated with viral isolation factors (region and time), especially the usages of thymine and cytosine. Despite the domination of adenine and thymine in the viral genome, we found that mutation pressure alters the overall codon usage pattern of ASFV, followed by selective forces from natural selection. Moreover, the nucleotide skew index at the gene level indicates that nucleotide usages influencing synonymous codon bias of ASFV are significantly correlated with viral protein hydropathy. Finally, evolutionary plasticity is proved to contribute to the weakness in synonymous codons with A- or T-end serving as optimal codons of ASFV, suggesting that fine-tuning translation selection plays a role in synonymous codon usages of ASFV for adapting host. Taken together, ASFV is subject to evolutionary dynamics on nucleotide selections and synonymous codon usage, and our detailed analysis offers deeper insights into the genetic characteristics of this newly emerging virus around the world.

Methylome Imputation by Methylation Patterns

DNA methylation is studied extensively for its relations with several biological processes such as transcriptional regulation. While methylation levels are usually estimated per cytosine or genomic region, additional information on methylation heterogeneity can be obtained when considering stretches of successive cytosines on the same reads; however, the majority of methylomes suffer from low coverage of genomic regions with sequencing depths enough for accurate estimation of methylation heterogeneity using existing methods. Here we describe a probabilistic-based imputation method that makes use of methylation information from neighboring sites to recover partially observed methylation patterns. Our method and software are proven to be faster and more accurate among all evaluated. Ultimately, our method allows for a more streamlined monitoring of epigenetic changes within cellular populations and their putative role in disease.

Influence of antidepressant treatment on SLC6A4 methylation in Korean patients with major depression

Genetic variation of the serotonin transporter gene (SLC6A4) has been suggested as potential mediator for antidepressant response in patients with depression. This study aimed to determine whether DNA methylation in SLC6A4 changes after antidepressant treatment and whether it affects treatment response in patients with depression. Overall, 221 Korean patients with depression completed 6 weeks of selective serotonin reuptake inhibitor (SSRI) monotherapy. DNA was extracted from venous blood pre- and post-treatment, and DNA methylation was analyzed using polymerase chain reaction. We used Wilcoxon's signed-rank test to verify the difference in methylation after treatment. Treatment response was assessed using the 17-item Hamilton Depression Rating Scale, and mRNA levels were quantified. After adjusting for relevant covariates, DNA methylation was significantly altered in specific CpG sites in SLC6A4 (p < .001 in CpG3, CpG4, and CpG5) following 6 weeks of treatment. Methylation change's magnitude (ΔDNA methylation) after drug treatment was not associated with treatment response or mRNA level change. SSRI antidepressants can influence SLC6A4 methylation in patients with depression. However, ΔDNA methylation at CpG3, CpG4, and CpG5 in SLC6A4 was not associated with treatment response. Future studies should investigate the integrative effect of other genetic variants and CpG methylation on gene transcription and antidepressant treatment response.

Identification of differentially methylated regions in rare diseases from a single-patient perspective

BACKGROUND

DNA methylation (5-mC) is being widely recognized as an alternative in the detection of sequence variants in the diagnosis of some rare neurodevelopmental and imprinting disorders. Identification of alterations in DNA methylation plays an important role in the diagnosis and understanding of the etiology of those disorders. Canonical pipelines for the detection of differentially methylated regions (DMRs) usually rely on inter-group (e.g., case versus control) comparisons. However, these tools might perform suboptimally in the context of rare diseases and multilocus imprinting disturbances due to small cohort sizes and inter-patient heterogeneity. Therefore, there is a need to provide a simple but statistically robust pipeline for scientists and clinicians to perform differential methylation analyses at the single patient level as well as to evaluate how parameter fine-tuning may affect differentially methylated region detection.

RESULT

We implemented an improved statistical method to detect differentially methylated regions in correlated datasets based on the Z-score and empirical Brown aggregation methods from a single-patient perspective. To accurately assess the predictive power of our method, we generated semi-simulated data using a public control population of 521 samples and investigated how the size of the control population, methylation difference, and region size affect DMR detection. In addition, we validated the detection of methylation events in patients suffering from rare multi-locus imprinting disturbance and evaluated how this method could complement existing tools in the context of clinical diagnosis.

CONCLUSION

In this study, we present a robust statistical method to perform differential methylation analysis at the single patient level and describe its optimal parameters to increase DMRs identification performance. Finally, we show its diagnostic utility when applied to rare disorders.

MC profiling: a novel approach to analyze DNA methylation heterogeneity in genome-wide bisulfite sequencing data

DNA methylation is an epigenetic mark implicated in crucial biological processes. Most of the knowledge about DNA methylation is based on bulk experiments, in which DNA methylation of genomic regions is reported as average methylation. However, average methylation does not inform on how methylated cytosines are distributed in each single DNA molecule. Here, we propose Methylation Class (MC) profiling as a genome-wide approach to the study of DNA methylation heterogeneity from bulk bisulfite sequencing experiments. The proposed approach is built on the concept of MCs, groups of DNA molecules sharing the same number of methylated cytosines. The relative abundances of MCs from sequencing reads incorporates the information on the average methylation, and directly informs on the methylation level of each molecule. By applying our approach to publicly available bisulfite-sequencing datasets, we individuated cell-to-cell differences as the prevalent contributor to methylation heterogeneity. Moreover, we individuated signatures of loci undergoing imprinting and X-inactivation, and highlighted differences between the two processes. When applying MC profiling to compare different conditions, we identified methylation changes occurring in regions with almost constant average methylation. Altogether, our results indicate that MC profiling can provide useful insights on the epigenetic status and its evolution at multiple genomic regions.

An evaluation of the genome-wide false positive rates of common methods for identifying differentially methylated regions using illumina methylation arrays

Differentially methylated regions (DMRs) are genomic regions with specific methylation patterns across multiple loci that are associated with a phenotype. We examined the genome-wide false positive (GFP) rates of five widely used DMR methods: comb-p, Bumphunter, DMRcate, mCSEA and coMethDMR using both Illumina HumanMethylation450 (450 K) and MethylationEPIC (EPIC) data and simulated continuous and dichotomous null phenotypes (i.e., generated independently of methylation data). coMethDMR provided well-controlled GFP rates (~5%) except when analysing skewed continuous phenotypes. DMRcate generally had well-controlled GFP rates when applied to 450 K data except for the skewed continuous phenotype and EPIC data only for the normally distributed continuous phenotype. GFP rates for mCSEA were at least 0.096 and comb-p yielded GFP rates above 0.34. Bumphunter had high GFP rates of at least 0.35 across conditions, reaching as high as 0.95. Analysis of the performance of these methods in specific regions of the genome found that regions with higher correlation across loci had higher regional false positive rates on average across methods. Based on the false positive rates, coMethDMR is the most recommended analysis method, and DMRcate had acceptable performance when analysing 450 K data. However, as both could display higher levels of FPs for skewed continuous distributions, a normalizing transformation of skewed continuous phenotypes is suggested. This study highlights the importance of genome-wide simulations when evaluating the performance of DMR-analysis methods.

DNA methylation loci identification for pan-cancer early-stage diagnosis and prognosis using a new distributed parallel partial least squares method

Aberrant methylation is one of the early detectable events in many tumors, which is very promising for pan-cancer early-stage diagnosis and prognosis. To efficiently analyze the big pan-cancer methylation data and to overcome the co-methylation phenomenon, a MapReduce-based distributed and parallel-designed partial least squares approach was proposed. The large-scale high-dimensional methylation data were first decomposed into distributed blocks according to their genome locations. A distributed and parallel data processing strategy was proposed based on the framework of MapReduce, and then latent variables were further extracted for each distributed block. A set of pan-cancer signatures through a differential co-expression network followed by statistical tests was further identified based on their gene expression profiles. In total, 15 TCGA and 3 GEO datasets were used as the training and testing data, respectively, to verify our method. As a result, 22,000 potential methylation loci were selected as highly related loci with early-stage pan-cancer diagnosis. Of these, 67 methylation loci were further identified as pan-cancer signatures considering their gene expression as well. The survival analysis as well as pathway enrichment analysis on them shows that not only these loci may serve as potential drug targets, but also the proposed method may serve as a uniform framework for signature identification with big data.

The Mutagenic Consequences of DNA Methylation within and across Generations

DNA methylation is an epigenetic modification with wide-ranging consequences across the life of an organism. This modification can be stable, persisting through development despite changing environmental conditions. However, in other contexts, DNA methylation can also be flexible, underlying organismal phenotypic plasticity. One underappreciated aspect of DNA methylation is that it is a potent mutagen; methylated cytosines mutate at a much faster rate than other genetic motifs. This mutagenic property of DNA methylation has been largely ignored in eco-evolutionary literature, despite its prevalence. Here, we explore how DNA methylation induced by environmental and other factors could promote mutation and lead to evolutionary change at a more rapid rate and in a more directed manner than through stochastic genetic mutations alone. We argue for future research on the evolutionary implications of DNA methylation driven mutations both within the lifetime of organisms, as well as across timescales.

The rs1001179 SNP and CpG methylation regulate catalase expression in chronic lymphocytic leukemia

Chronic lymphocytic leukemia (CLL) is an incurable disease characterized by an extremely variable clinical course. We have recently shown that high catalase (CAT) expression identifies patients with an aggressive clinical course. Elucidating mechanisms regulating CAT expression in CLL is preeminent to understand disease mechanisms and develop strategies for improving its clinical management. In this study, we investigated the role of the CAT promoter rs1001179 single nucleotide polymorphism (SNP) and of the CpG Island II methylation encompassing this SNP in the regulation of CAT expression in CLL. Leukemic cells harboring the rs1001179 SNP T allele exhibited a significantly higher CAT expression compared with cells bearing the CC genotype. CAT promoter harboring the T -but not C- allele was accessible to ETS-1 and GR-β transcription factors. Moreover, CLL cells exhibited lower methylation levels than normal B cells, in line with the higher CAT mRNA and protein expressed by CLL in comparison with normal B cells. Methylation levels at specific CpG sites negatively correlated with CAT levels in CLL cells. Inhibition of methyltransferase activity induced a significant increase in CAT levels, thus functionally validating the role of CpG methylation in regulating CAT expression in CLL. Finally, the CT/TT genotypes were associated with lower methylation and higher CAT levels, suggesting that the rs1001179 T allele and CpG methylation may interact in regulating CAT expression in CLL. This study identifies genetic and epigenetic mechanisms underlying differential expression of CAT, which could be of crucial relevance for the development of therapies targeting redox regulatory pathways in CLL.

DNA Methylation Profiles of GAD1 in Human Cerebral Organoids of Autism Indicate Disrupted Epigenetic Regulation during Early Development

DNA methylation profiling has become a promising approach towards identifying biomarkers of neuropsychiatric disorders including autism spectrum disorder (ASD). Epigenetic markers capture genetic risk factors and diverse exogenous and endogenous factors, including environmental risk factors and complex disease pathologies. We analysed the differential methylation profile of a regulatory region of the GAD1 gene using cerebral organoids generated from induced pluripotent stem cells (iPSCs) from adults with a diagnosis of ASD and from age- and gender-matched healthy individuals. Both groups showed high levels of methylation across the majority of CpG sites within the profiled GAD1 region of interest. The ASD group exhibited a higher number of unique DNA methylation patterns compared to controls and an increased CpG-wise variance. We detected six differentially methylated CpG sites in ASD, three of which reside within a methylation-dependent transcription factor binding site. In ASD, GAD1 is subject to differential methylation patterns that may not only influence its expression, but may also indicate variable epigenetic regulation among cells.

Evolutionary dynamics of codon usages for peste des petits ruminants virus

Peste des petits ruminants virus (PPRV) is an important agent of contagious, acute and febrile viral diseases in small ruminants, while its evolutionary dynamics related to codon usage are still lacking. Herein, we adopted information entropy, the relative synonymous codon usage values and similarity indexes and codon adaptation index to analyze the viral genetic features for 45 available whole genomes of PPRV. Some universal, lineage-specific, and gene-specific genetic features presented by synonymous codon usages of the six genes of PPRV that encode N, P, M, F, H and L proteins reflected evolutionary plasticity and independence. The high adaptation of PPRV to hosts at codon usages reflected high viral gene expression, but some synonymous codons that are rare in the hosts were selected in high frequencies in the viral genes. Another obvious genetic feature was that the synonymous codons containing CpG dinucleotides had weak tendencies to be selected in viral genes. The synonymous codon usage patterns of PPRV isolated during 2007–2008 and 2013–2014 in China displayed independent evolutionary pathway, although the overall codon usage patterns of these PPRV strains matched the universal codon usage patterns of lineage IV. According to the interplay between nucleotide and synonymous codon usages of the six genes of PPRV, the evolutionary dynamics including mutation pressure and natural selection determined the viral survival and fitness to its host.

The prediction of tumor and normal tissues based on the DNA methylation values of ten key sites

Abnormal DNA methylation can alter the gene expression to promote or inhibit tumorigenesis in colon adenocarcinoma (COAD). However, the finding important genes and key sites of abnormal DNA methylation which result in the occurrence of COAD is still an eventful task. Here, we studied the effects of DNA methylation in the 12 types of genomic features on the changes of gene expression in COAD, the 10 important COAD-related genes and the key abnormal DNA methylation sites were identified. The effects of important genes on the prognosis were verified by survival analysis. Moreover, it was shown that the important genes were participated in cancer pathways and were hub genes in a co-expression network. Based on the DNA methylation levels in the ten sites, the least diversity increment algorithm for predicting tumor tissues and normal tissues in seventeen cancer types are proposed. The better results are obtained in jackknife test. For example, the predictive accuracies are 94.17 %, 91.28 %, 89.04 % and 88.89 %, respectively, for COAD, rectum adenocarcinoma, pancreatic adenocarcinoma and cholangiocarcinoma. Finally, by computing enrichment score of infiltrating immunocytes and the activity of immune pathways, we found that the genes are highly correlated with immune microenvironment.

Identification of Endocannabinoid Predictors of Treatment Outcomes in Major Depressive Disorder: A Secondary Analysis of the First Canadian Biomarker Integration Network in Depression (CAN-BIND 1) Study

INTRODUCTION

An increasing number of studies are examining the link between the endocannabinoidome and major depressive disorder (MDD). We conducted an exploratory analysis of this system to identify potential markers of treatment outcomes.

METHODS

The dataset of the Canadian Biomarker Integration Network in Depression-1 study, consisting of 180 patients with MDD treated for eight weeks with escitalopram followed by eight weeks with escitalopram alone or augmented with aripiprazole was analyzed. Association between response Montgomery-Asberg Depression Rating Scale (MADRS; score reduction≥50%) or remission (MADRS score≤10) at weeks 8 and 16 and single nucleotide polymorphisms (SNPs), methylation, and mRNA levels of 33 endocannabinoid markers were examined. A standard genome-wide association studies protocol was used for identifying SNPs, and logistic regression was used to assess methylation and mRNA levels.

RESULTS

Lower methylation of CpG islands of the diacylglycerol lipase alpha gene (DAGLA) was associated with non-remission at week 16 (DAGLA; OR=0.337, p<0.003, q=0.050). Methylation of DAGLA was correlated with improvement in Clinical Global Impression (p=0.026), Quick Inventory of Depressive Symptomatology (p=0.010), and Snaith-Hamilton Pleasure scales (p=0.028). We did not find any association between SNPs or mRNA levels and treatment outcomes.

DISCUSSION

Methylation of DAGLA is a promising candidate as a marker of treatment outcomes for MDD and needs to be explored further.

Random field modeling of multi-trait multi-locus association for detecting methylation quantitative trait loci

MOTIVATION

CpG sites within the same genomic region often share similar methylation patterns and tend to be co-regulated by multiple genetic variants that may interact with one another.

RESULTS

We propose a multi-trait methylation random field (multi-MRF) method to evaluate the joint association between a set of CpG sites and a set of genetic variants. The proposed method has several advantages. First, it is a multi-trait method that allows flexible correlation structures between neighboring CpG sites (e.g. distance-based correlation). Second, it is also a multi-locus method that integrates the effect of multiple common and rare genetic variants. Third, it models the methylation traits with a beta distribution to characterize their bimodal and interval properties. Through simulations, we demonstrated that the proposed method had improved power over some existing methods under various disease scenarios. We further illustrated the proposed method via an application to a study of congenital heart defects (CHDs) with 83 cardiac tissue samples. Our results suggested that gene BACE2, a methylation quantitative trait locus (QTL) candidate, colocalized with expression QTLs in artery tibial and harbored genetic variants with nominal significant associations in two genome-wide association studies of CHD.

AVAILABILITY AND IMPLEMENTATION

https://github.com/chenlyu2656/Multi-MRF.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

BSImp: Imputing Partially Observed Methylation Patterns for Evaluating Methylation Heterogeneity

DNA methylation is one of the most studied epigenetic modifications that has applications ranging from transcriptional regulation to aging, and can be assessed by bisulfite sequencing (BS-seq) or enzymatic methyl sequencing (EM-seq) at single base-pair resolution. The permutations of methylation statuses given by aligned reads reflect the methylation patterns of individual cells. These patterns at specific genomic locations are sought to be indicative of cellular heterogeneity within a cellular population, which are predictive of developments and diseases; therefore, methylation heterogeneity has potentials in early detection of these changes. Computational methods have been developed to assess methylation heterogeneity using methylation patterns formed by four consecutive CpGs, but the nature of shotgun sequencing often give partially observed patterns, which makes very limited data available for downstream analysis. While many programs are developed to impute genome-wide methylation levels, currently there is only one method developed for recovering partially observed methylation patterns; however, the program needs lots of data to train and cannot be used directly; therefore, we developed a probabilistic-based imputation method that uses information from neighbouring sites to recover partially observed methylation patterns speedily. It is demonstrated to allow for the evaluation of methylation heterogeneity at 15% more regions genome-wide with high accuracy for data with moderate depth. To make it more user-friendly we also provide a computational pipeline for genome-screening, which can be used in both evaluating methylation levels and profiling methylation patterns genomewide for all cytosine contexts, which is the first of its kind. Our method allows for accurate estimation of methylation levels and makes evaluating methylation heterogeneity available for much more data with reasonable coverage, which has important implications in using methylation heterogeneity for monitoring changes within the cellular populations that were impossible to detect for the assessment of development and diseases.

Modeling dependency structures in 450k DNA methylation data

MOTIVATION

DNA methylation has been shown to be spatially dependent across chromosomes. Previous studies have focused on the influence of genomic context on the dependency structure, while not considering differences in dependency structure between individuals.

RESULTS

We modeled spatial dependency with a flexible framework to quantify the dependency structure, focusing on inter-individual differences by exploring the association between dependency parameters and technical and biological variables. The model was applied to a subset of the Finnish Twin Cohort study (N = 1611 individuals). The estimates of the dependency parameters varied considerably across individuals, but were generally consistent across chromosomes within individuals. The variation in dependency parameters was associated with bisulfite conversion plate, zygosity, sex and age. The age differences presumably reflect accumulated environmental exposures and/or accumulated small methylation differences caused by stochastic mitotic events, establishing recognizable, individual patterns more strongly seen in older individuals.

AVAILABILITY AND IMPLEMENTATION

The twin dataset used in the current study are located in the Biobank of the National Institute for Health and Welfare, Finland. All the biobanked data are publicly available for use by qualified researchers following a standardized application procedure (https://thl.fi/en/web/thl-biobank/for-researchers). A R-script for fitting the dependency structure to publicly available DNA methylation data with the software used in this article is provided in supplementary data.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

MethylSPWNet and MethylCapsNet: Biologically Motivated Organization of DNAm Neural Networks, Inspired by Capsule Networks

DNA methylation (DNAm) alterations have been heavily implicated in carcinogenesis and the pathophysiology of diseases through upstream regulation of gene expression. DNAm deep-learning approaches are able to capture features associated with aging, cell type, and disease progression, but lack incorporation of prior biological knowledge. Here, we present modular, user-friendly deep-learning methodology and software, MethylCapsNet and MethylSPWNet, that group CpGs into biologically relevant capsules-such as gene promoter context, CpG island relationship, or user-defined groupings-and relate them to diagnostic and prognostic outcomes. We demonstrate these models' utility on 3,897 individuals in the classification of central nervous system (CNS) tumors. MethylCapsNet and MethylSPWNet provide an opportunity to increase DNAm deep-learning analyses' interpretability by enabling a flexible organization of DNAm data into biologically relevant capsules.

Hypomethylation of AHRR (cg05575921) Is Related to Smoking Status in the Mexican Mestizo Population

Tobacco smoking results in a multifactorial disease involving environmental and genetic factors; epigenome-wide association studies (EWAS) show changes in DNA methylation levels due to cigarette consumption, partially reversible upon tobacco smoking cessation. Therefore, methylation levels could predict smoking status. This study aimed to evaluate the DNA methylation level of cg05575921 (AHRR) and cg23771366 (PRSS23) and their correlation with lung function variables, cigarette consumption, and nicotine addiction in the Mexican smoking population. We included 114 non-smokers (NS) and 102 current tobacco smokers (TS); we then further subclassified them as heavy smokers (HS) (n = 53) and light smokers (LS) (n = 49). We used restriction enzymes (MspI/HpaII) and qPCR to determine the DNA methylation level. We observed significant hypomethylation of cg05575921 in smokers compared to NS (p = 0.003); further analysis found a difference between HS and NS (p = 0.02). We did not observe differences between other groups or a positive correlation between methylation levels and age, BMI, cigarette consumption, nicotine addiction, or lung function. In conclusion, the cg05575921 site of AHRR is significantly hypomethylated in Mexican smokers, especially in HS (≥20 cigarettes per day).

Disentangling tumorigenesis-associated DNA methylation changes in colorectal tissues from those associated with ageing

Physiological ageing and tumorigenesis are both associated with epigenomic alterations in human tissue cells, the most extensively investigated of which entails de novo cytosine methylation (i.e., hypermethylation) within the CpG dinucleotides of CpG islands. Genomic regions that become hypermethylated during tumorigenesis are generally believed to overlap regions that acquire methylation in normal tissues as an effect of ageing. To define the extension of this overlap, we analysed the DNA methylomes of 48 large-bowel tissue samples taken from women of different ages during screening colonoscopy: 18 paired samples of normal and lesional tissues from donors harbouring a precancerous lesion and 12 samples of normal mucosa from tumour-free donors. Each sample was subjected to targeted, genome-wide bisulphite sequencing of ~2.5% of the genome, including all CpG islands. In terms of both its magnitude and extension along the chromatin, tumour-associated DNA hypermethylation in these regions was much more conspicuous than that observed in the normal mucosal samples from older (vs. younger) tumour-free donors. 83% of the ageing-associated hypermethylated regions (n = 2501) coincided with hypermethylated regions observed in tumour samples. However, 86% of the regions displaying hypermethylation in precancerous lesions (n = 16,772) showed no methylation changes in the ageing normal mucosa. The tumour-specificity of this latter hypermethylation was validated using published sets of data on DNA methylation in normal and neoplastic colon tissues. This extensive set of genomic regions displaying tumour-specific hypermethylation represents a rich vein of putative biomarkers for the early, non-invasive detection of colorectal tumours in women of all ages.

Genome-wide detection of cytosine methylation by single molecule real-time sequencing

Single molecule real-time (SMRT) sequencing theoretically offers the opportunity to directly assess certain base modifications of native DNA molecules without any prior chemical/enzymatic conversions and PCR amplification, using kinetic signals of a DNA polymerase. However, the kinetic signal changes caused by 5mC modification are extremely subtle. Hence, the robust genome-wide measurement of 5mC modification has not been achieved. We enhanced 5mC detection using SMRT sequencing by holistically analyzing kinetic signals of a DNA polymerase and sequence context for every base within a measurement window. We employed a convolutional neural network to train a methylation classification model, leading to genome-wide 5mC detection. The sensitivity and specificity reached 90% and 94%, with a 99% correlation of overall methylation level with bisulfite sequencing.

5-Methylcytosine (5mC) is an important type of epigenetic modification. Bisulfite sequencing (BS-seq) has limitations, such as severe DNA degradation. Using single molecule real-time sequencing, we developed a methodology to directly examine 5mC. This approach holistically examined kinetic signals of a DNA polymerase (including interpulse duration and pulse width) and sequence context for every nucleotide within a measurement window, termed the holistic kinetic (HK) model. The measurement window of each analyzed double-stranded DNA molecule comprised 21 nucleotides with a cytosine in a CpG site in the center. We used amplified DNA (unmethylated) and M.SssI-treated DNA (methylated) (M.SssI being a CpG methyltransferase) to train a convolutional neural network. The area under the curve for differentiating methylation states using such samples was up to 0.97. The sensitivity and specificity for genome-wide 5mC detection at single-base resolution reached 90% and 94%, respectively. The HK model was then tested on human–mouse hybrid fragments in which each member of the hybrid had a different methylation status. The model was also tested on human genomic DNA molecules extracted from various biological samples, such as buffy coat, placental, and tumoral tissues. The overall methylation levels deduced by the HK model were well correlated with those by BS-seq (r = 0.99; P < 0.0001) and allowed the measurement of allele-specific methylation patterns in imprinted genes. Taken together, this methodology has provided a system for simultaneous genome-wide genetic and epigenetic analyses.

Genetic impacts on DNA methylation: research findings and future perspectives

Multiple recent studies highlight that genetic variants can have strong impacts on a significant proportion of the human DNA methylome. Methylation quantitative trait loci, or meQTLs, allow for the exploration of biological mechanisms that underlie complex human phenotypes, with potential insights for human disease onset and progression. In this review, we summarize recent milestones in characterizing the human genetic basis of DNA methylation variation over the last decade, including heritability findings and genome-wide identification of meQTLs. We also discuss challenges in this field and future areas of research geared to generate insights into molecular processes underlying human complex traits.

The Detection of Cancer Epigenetic Traces in Cell-Free DNA

Nucleic acid fragments found in blood circulation originate mostly from dying cells and carry signs pointing to specific features of the parental cell types. Deciphering these clues may be transformative for numerous research and clinical applications but strongly depends on the development and implementation of robust analytical methods. Remarkable progress has been achieved in the reliable detection of sequence alterations in cell-free DNA while decoding epigenetic information from methylation and fragmentation patterns requires more sophisticated approaches. This review discusses the currently available strategies for detecting and analyzing the epigenetic marks in the liquid biopsies.

Identification of methylation changes associated with positive and negative growth deviance in Gambian infants using a targeted methyl sequencing approach of genomic DNA

Low birthweight and reduced height gain during infancy (stunting) may arise at least in part from adverse early life environments that trigger epigenetic reprogramming that may favor survival. We examined differential DNA methylation patterns using targeted methyl sequencing of regions regulating gene activity in groups of rural Gambian infants: (a) low and high birthweight (DNA from cord blood (n = 16 and n = 20, respectively), from placental trophoblast tissue (n = 21 and n = 20, respectively), and DNA from peripheral blood collected from infants at 12 months of age (n = 23 and n = 17, respectively)), and, (b) the top 10% showing rapid postnatal length gain (high, n = 20) and the bottom 10% showing slow postnatal length gain (low, n = 20) based on z score change between birth and 12 months of age (LAZ) (DNA from peripheral blood collected from infants at 12 months of age). Using BiSeq analysis to identify significant methylation marks, for birthweight, four differentially methylated regions (DMRs) were identified in trophoblast DNA, compared to 68 DMRs in cord blood DNA, and 54 DMRs in 12‐month peripheral blood DNA. Twenty‐five DMRs were observed to be associated with high and low length for age (LAZ) at 12 months. With the exception of five loci (associated with two different genes), there was no overlap between these groups of methylation marks. Of the 194 CpG methylation marks contained within DMRs, 106 were located to defined gene regulatory elements (promoters, CTCF‐binding sites, transcription factor‐binding sites, and enhancers), 58 to gene bodies (introns or exons), and 30 to intergenic DNA. Distinct methylation patterns associated with birthweight between comparison groups were observed in DNA collected at birth (at the end of intrauterine growth window) compared to those established by 12 months (near the infancy/childhood growth transition). The longitudinal differences in methylation patterns may arise from methylation adjustments, changes in cellular composition of blood or both that continue during the critical postnatal growth period, and in response to early nutritional and infectious environmental exposures with impacts on growth and longer‐term health outcomes.

Cell lineage-specific methylome and genome alterations in gout

In this study, we examined data from 69 gout patients and 1,455 non-gout controls using a MethylationEPIC BeadChip assay and Illumina HiSeq platform to identify lineage-specific epigenetic alterations and associated genetic factors that contributed to gouty inflammation. Cell lineage-specific differentially methylated sites were identified using CellDMC after adjusting for sex, age, alcohol drinking, smoking status, and smoking history (total pack-years). Different cell lineages displayed distinct differential methylation. Ingenuity Pathway Analysis and NetworkAnalyst indicated that many differential methylated sites were associated with interleukin-1β expression in monocytes. On the UCSC Genome Browser and WashU Epigenome Browser, metabolic trait, cis-methylation quantitative trait loci, genetic, and functional annotation analyses identified nine methylation loci located in interleukin-1β-regulating genes (PRKCZ, CIDEC, VDAC1, CPT1A, BIRC2, BRCA1, STK11, and NLRP12) that were associated specifically with gouty inflammation. All nine sites mapped to active regulatory elements in monocytes. MoLoTool and ReMap analyses indicated that the nine methylation loci overlapped with binding sites of several transcription factors that regulated interleukin-1β production and gouty inflammation. Decreases in PRKCZ and STK11 methylation were also associated with higher numbers of first-degree relatives who also had gout. The gouty-inflammation specific methylome and genome alterations could potentially aid in the identification of novel therapeutic targets.

CaMelia: imputation in single-cell methylomes based on local similarities between cells

MOTIVATION

Single-cell DNA methylation sequencing detects methylation levels with single-cell resolution, while this technology is upgrading our understanding of the regulation of gene expression through epigenetic modifications. Meanwhile, almost all current technologies suffer from the inherent problem of detecting low coverage of the number of CpGs. Therefore, addressing the inherent sparsity of raw data is essential for quantitative analysis of the whole genome.

RESULTS

Here, we reported CaMelia, a CatBoost gradient boosting method for predicting the missing methylation states based on the locally paired similarity of intercellular methylation patterns. On real single-cell methylation data sets, CaMelia yielded significant imputation performance gains over previous methods. Furthermore, applying the imputed data to the downstream analysis of cell-type identification, we found that CaMelia helped to discover more intercellular differentially methylated loci that were masked by the sparsity in raw data, and the clustering results demonstrated that CaMelia could preserve cell-cell relationships and improve the identification of cell types and cell subpopulations.

AVAILABILITY

Python code is available at https://github.com/JxTang-bioinformatics/CaMelia.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Short history of 5-methylcytosine: from discovery to clinical applications

Covalent modifications of nucleotides in genetic material have been known from the beginning of the last century. Currently, one of those modifications referred to as DNA methylation, is impacting personalised medicine both as a treatment target and a biomarker source for clinical disease management. In this short review, we describe landmark discoveries that led to the elucidation of the DNA methylation importance in the cell's physiology and clarification of its role as one of the major processes in disease pathology. We also describe turning points in the development of methodologies to study this modification, which ultimately resulted in the development of in-vitro diagnostic kits targeting disease related DNA methylation changes as biomarkers.

Environmental and socio-cultural impacts on global DNA methylation in the indigenous Huichol population of Nayarit, Mexico

Alterations of global DNA methylation have been evaluated in several studies worldwide; however, Long Interspersed Nuclear Elements-1 (LINE-1) methylation in genetically conserved populations such as indigenous communities have not, to our knowledge, been reported. The aim of this study was to evaluate the relationship between LINE-1 methylation patterns and factors such as pesticide exposure and socio-cultural characteristics in the Indigenous Huichol Population of Nayarit, Mexico. A cross-sectional study was conducted in 140 Huichol indigenous individuals. A structured questionnaire was used to determine general and anthropometric characteristics, diet, harmful habits, and pesticide exposure. DNA methylation was determined by pyrosequencing of bisulfite-treated DNA. A lower level of LINE-1 methylation was found in the indigenous population when compared to a Mestizo population previously studied by our group. This difference might be due to the influence of the genetic admixture and differing dietary and lifestyle habits. The males in the indigenous population exhibited increased LINE-1 methylation in comparison to the females. Sex and alcohol consumption showed positive associations with LINE-1 methylation, while weight, current work in the field, current pesticide usage, and folate intake exhibited negative associations with LINE-1 methylation. The results suggest that ethnicity, as well as other internal and environmental factors, might influence LINE-1 methylation.

Epigenetic profiling of Italian patients identified methylation sites associated with hereditary transthyretin amyloidosis

Hereditary transthyretin (TTR) amyloidosis (hATTR) is a rare life-threatening disorder caused by amyloidogenic coding mutations located in TTR gene. To understand the high phenotypic variability observed among carriers of TTR disease-causing mutations, we conducted an epigenome-wide association study (EWAS) assessing more than 700,000 methylation sites and testing epigenetic difference of TTR coding mutation carriers vs. non-carriers. We observed a significant methylation change at cg09097335 site located in Beta-secretase 2 (BACE2) gene (standardized regression coefficient = -0.60, p = 6.26 × 10-8). This gene is involved in a protein interaction network enriched for biological processes and molecular pathways related to amyloid-beta metabolism (Gene Ontology: 0050435, q = 0.007), amyloid fiber formation (Reactome HSA-977225, q = 0.008), and Alzheimer's disease (KEGG hsa05010, q = 2.2 × 10-4). Additionally, TTR and BACE2 share APP (amyloid-beta precursor protein) as a validated protein interactor. Within TTR gene region, we observed that Val30Met disrupts a methylation site, cg13139646, causing a drastic hypomethylation in carriers of this amyloidogenic mutation (standardized regression coefficient = -2.18, p = 3.34 × 10-11). Cg13139646 showed co-methylation with cg19203115 (Pearson's r2 = 0.32), which showed significant epigenetic differences between symptomatic and asymptomatic carriers of amyloidogenic mutations (standardized regression coefficient = -0.56, p = 8.6 × 10-4). In conclusion, we provide novel insights related to the molecular mechanisms involved in the complex heterogeneity of hATTR, highlighting the role of epigenetic regulation in this rare disorder.

Modeling DNA Methylation Profiles through a Dynamic Equilibrium between Methylation and Demethylation

DNA methylation is a heritable epigenetic mark that plays a key role in regulating gene expression. Mathematical modeling has been extensively applied to unravel the regulatory mechanisms of this process. In this study, we aimed to investigate DNA methylation by performing a high-depth analysis of particular loci, and by subsequent modeling of the experimental results. In particular, we performed an in-deep DNA methylation profiling of two genomic loci surrounding the transcription start site of the D-Aspartate Oxidase and the D-Serine Oxidase genes in different samples (n = 51). We found evidence of cell-to-cell differences in DNA methylation status. However, these cell differences were maintained between different individuals, which indeed showed very similar DNA methylation profiles. Therefore, we hypothesized that the observed pattern of DNA methylation was the result of a dynamic balance between DNA methylation and demethylation, and that this balance was identical between individuals. We hence developed a simple mathematical model to test this hypothesis. Our model reliably captured the characteristics of the experimental data, suggesting that DNA methylation and demethylation work together in determining the methylation state of a locus. Furthermore, our model suggested that the methylation status of neighboring cytosines plays an important role in this balance.

Epigenome-wide association studies in Alzheimer's disease; achievements and challenges

Alzheimer's disease (AD) represents a devastating progressive neurodegenerative disease with a complex pathophysiology, affecting millions of people worldwide. Recent epigenome-wide association studies suggest a key role for epigenetic mechanisms in its development and course. Despite the fact that current evidence on the role of epigenetic dysregulation in aging and AD is convincing, the pioneering field of neuroepigenetics is still facing many challenges that need to be addressed to fundamentally increase our understanding about the underlying mechanisms of this neurodegenerative disorder. This perspective paper describes the current state of play for epigenetic research into AD and discusses how new methodological advances in the field of epigenetics and related data science disciplines could further spur the development of novel therapeutic agents and biomarker assays.

Distinct viral reservoirs in individuals with spontaneous control of HIV-1

Sustained, drug-free control of HIV-1 replication is naturally achieved in less than 0.5% of infected individuals (here termed 'elite controllers'), despite the presence of a replication-competent viral reservoir1. Inducing such an ability to spontaneously maintain undetectable plasma viraemia is a major objective of HIV-1 cure research, but the characteristics of proviral reservoirs in elite controllers remain to be determined. Here, using next-generation sequencing of near-full-length single HIV-1 genomes and corresponding chromosomal integration sites, we show that the proviral reservoirs of elite controllers frequently consist of oligoclonal to near-monoclonal clusters of intact proviral sequences. In contrast to individuals treated with long-term antiretroviral therapy, intact proviral sequences from elite controllers were integrated at highly distinct sites in the human genome and were preferentially located in centromeric satellite DNA or in Krüppel-associated box domain-containing zinc finger genes on chromosome 19, both of which are associated with heterochromatin features. Moreover, the integration sites of intact proviral sequences from elite controllers showed an increased distance to transcriptional start sites and accessible chromatin of the host genome and were enriched in repressive chromatin marks. These data suggest that a distinct configuration of the proviral reservoir represents a structural correlate of natural viral control, and that the quality, rather than the quantity, of viral reservoirs can be an important distinguishing feature for a functional cure of HIV-1 infection. Moreover, in one elite controller, we were unable to detect intact proviral sequences despite analysing more than 1.5 billion peripheral blood mononuclear cells, which raises the possibility that a sterilizing cure of HIV-1 infection, which has previously been observed only following allogeneic haematopoietic stem cell transplantation2,3, may be feasible in rare instances.

Gene-methylation interactions: discovering region-wise DNA methylation levels that modify SNP-associated disease risk

BACKGROUND

Current technology allows rapid assessment of DNA sequences and methylation levels at a single-site resolution for hundreds of thousands of sites in the human genome, in thousands of individuals simultaneously. This has led to an increase in epigenome-wide association studies (EWAS) of complex traits, particularly those that are poorly explained by previous genome-wide association studies (GWAS). However, the genome and epigenome are intertwined, e.g., DNA methylation is known to affect gene expression through, for example, genomic imprinting. There is thus a need to go beyond single-omics data analyses and develop interaction models that allow a meaningful combination of information from EWAS and GWAS.

RESULTS

We present two new methods for genetic association analyses that treat offspring DNA methylation levels as environmental exposure. Our approach searches for statistical interactions between SNP alleles and DNA methylation (G ×Me) and between parent-of-origin effects and DNA methylation (PoO ×Me), using case-parent triads or dyads. We use summarized methylation levels over nearby genomic region to ease biological interpretation. The methods were tested on a dataset of parent-offspring dyads, with EWAS data on the offspring. Our results showed that methylation levels around a SNP can significantly alter the estimated relative risk. Moreover, we show how a control dataset can identify false positives.

CONCLUSIONS

The new methods, G ×Me and PoO ×Me, integrate DNA methylation in the assessment of genetic relative risks and thus enable a more comprehensive biological interpretation of genome-wide scans. Moreover, our strategy of condensing DNA methylation levels within regions helps overcome specific disadvantages of using sparse chip-based measurements. The methods are implemented in the freely available R package Haplin ( https://cran.r-project.org/package=Haplin ), enabling fast scans of multi-omics datasets.

Systemic Investigation of Promoter-wide Methylome and Genome Variations in Gout

Current knowledge of gout centers on hyperuricemia. Relatively little is known regarding the pathogenesis of gouty inflammation. To investigate the epigenetic background of gouty inflammation independent of hyperuricemia and its relationship to genetics, 69 gout patients and 1455 non-gout controls were included. Promoter-wide methylation was profiled with EPIC array. Whole-genome sequencing data were included for genetic and methylation quantitative trait loci (meQTL) analyses and causal inference tests. Identified loci were subjected to co-methylation analysis and functional localization with DNase hypersensitivity and histone marks analysis. An expression database was queried to clarify biologic functions of identified loci. A transcription factor dataset was integrated to identify transcription factors coordinating respective expression. In total, seven CpG loci involved in interleukin-1β production survived genetic/meQTL analyses, or causal inference tests. None had a significant relationship with various metabolic traits. Additional analysis suggested gouty inflammation, instead of hyperuricemia, provides the link between these CpG sites and gout. Six (PGGT1B, INSIG1, ANGPTL2, JNK1, UBAP1, and RAPTOR) were novel genes in the field of gout. One (CNTN5) was previously associated with gouty inflammation. Transcription factor mapping identified several potential transcription factors implicated in the link between differential methylation, interleukin-1β production, and gouty inflammation. In conclusion, this study revealed several novel genes specific to gouty inflammation and provided enhanced insight into the biological basis of gouty inflammation.

Nutritional Epigenomics and Age-Related Disease

Recent advances in epigenetic research have enabled the development of epigenetic clocks, which have greatly enhanced our ability to investigate molecular processes that contribute to aging and age-related disease. These biomarkers offer the potential to measure the effect of environmental exposures linked to dynamic changes in DNA methylation, including nutrients, as factors in age-related disease. They also offer a compelling insight into how imbalances in the supply of nutrients, particularly B-vitamins, or polymorphisms in regulatory enzymes involved in 1-carbon metabolism, the key pathway that supplies methyl groups for epigenetic reactions, may influence epigenetic age and interindividual disease susceptibility. Evidence from recent studies is critically reviewed, focusing on the significant contribution of the epigenetic clock to nutritional epigenomics and its impact on health outcomes and age-related disease. Further longitudinal studies and randomized nutritional interventions are required to advance the field.

DAMEfinder: a method to detect differential allele-specific methylation

Background

DNA methylation is a highly studied epigenetic signature that is associated with regulation of gene expression, whereby genes with high levels of promoter methylation are generally repressed. Genomic imprinting occurs when one of the parental alleles is methylated, i.e., when there is inherited allele-specific methylation (ASM). A special case of imprinting occurs during X chromosome inactivation in females, where one of the two X chromosomes is silenced, to achieve dosage compensation between the sexes. Another more widespread form of ASM is sequence dependent (SD-ASM), where ASM is linked to a nearby heterozygous single nucleotide polymorphism (SNP).

Results

We developed a method to screen for genomic regions that exhibit loss or gain of ASM in samples from two conditions (treatments, diseases, etc.). The method relies on the availability of bisulfite sequencing data from multiple samples of the two conditions. We leverage other established computational methods to screen for these regions within a new R package called DAMEfinder. It calculates an ASM score for all CpG sites or pairs in the genome of each sample, and then quantifies the change in ASM between conditions. It then clusters nearby CpG sites with consistent change into regions. In the absence of SNP information, our method relies only on reads to quantify ASM. This novel ASM score compares favorably to current methods that also screen for ASM. Not only does it easily discern between imprinted and non-imprinted regions, but also females from males based on X chromosome inactivation. We also applied DAMEfinder to a colorectal cancer dataset and observed that colorectal cancer subtypes are distinguishable according to their ASM signature. We also re-discover known cases of loss of imprinting.

Conclusion

We have designed DAMEfinder to detect regions of differential ASM (DAMEs), which is a more refined definition of differential methylation, and can therefore help in breaking down the complexity of DNA methylation and its influence in development and disease.

The DNA hypermethylation phenotype of colorectal cancer liver metastases resembles that of the primary colorectal cancers

BACKGROUND

Identifying molecular differences between primary and metastatic colorectal cancers-now possible with the aid of omics technologies-can improve our understanding of the biological mechanisms of cancer progression and facilitate the discovery of novel treatments for late-stage cancer. We compared the DNA methylomes of primary colorectal cancers (CRCs) and CRC metastases to the liver. Laser microdissection was used to obtain epithelial tissue (10 to 25 × 10⁶ μm²) from sections of fresh-frozen samples of primary CRCs (n = 6), CRC liver metastases (n = 12), and normal colon mucosa (n = 3). DNA extracted from tissues was enriched for methylated sequences with a methylCpG binding domain (MBD) polypeptide-based protocol and subjected to deep sequencing. The performance of this protocol was compared with that of targeted enrichment for bisulfite sequencing used in a previous study of ours.

RESULTS

MBD enrichment captured a total of 322,551 genomic regions (249.5 Mb or ~ 7.8% of the human genome), which included over seven million CpG sites. A few of these regions were differentially methylated at an expected false discovery rate (FDR) of 5% in neoplastic tissues (primaries: 0.67%, i.e., 2155 regions containing 279,441 CpG sites; liver metastases: 1%, i.e., 3223 regions containing 312,723 CpG sites) as compared with normal mucosa samples. Most of the differentially methylated regions (DMRs; 94% in primaries; 70% in metastases) were hypermethylated, and almost 80% of these (1882 of 2396) were present in both lesion types. At 5% FDR, no DMRs were detected in liver metastases vs. primary CRC. However, short regions of low-magnitude hypomethylation were frequent in metastases but rare in primaries. Hypermethylated DMRs were far more abundant in sequences classified as intragenic, gene-regulatory, or CpG shelves-shores-island segments, whereas hypomethylated DMRs were equally represented in extragenic (mainly, open-sea) and intragenic (mainly, gene bodies) sequences of the genome. Compared with targeted enrichment, MBD capture provided a better picture of the extension of CRC-associated DNA hypermethylation but was less powerful for identifying hypomethylation.

CONCLUSIONS

Our findings demonstrate that the hypermethylation phenotype in CRC liver metastases remains similar to that of the primary tumor, whereas CRC-associated DNA hypomethylation probably undergoes further progression after the cancer cells have migrated to the liver.