MethLAB: a graphical user interface package for the analysis of array-based DNA methylation data

Controlling Batch Effect in Epigenome-Wide Association Study

Methylation data, similar to other omics data, is susceptible to various technical issues that are potentially associated with unexplained or unrelated factors. Any difference in the measurement of DNA methylation, such as laboratory operation and sequencing platform, may lead to batch effects. With the accumulation of large-scale omics data, scientists are making joint efforts to generate and analyze omics data to answer various scientific questions. However, batch effects are inevitable in practice, and careful adjustment is needed. Multiple statistical methods for controlling bias and inflation between batches have been developed either by correcting based on known batch factors or by estimating directly from the output data. In this chapter, we will review and demonstrate several popular methods for batch effect correction and make practical recommendations in epigenome-wide association studies (EWAS).

Filtering High-Dimensional Methylation Marks With Extremely Small Sample Size: An Application to Gastric Cancer Data

DNA methylations in critical regions are highly involved in cancer pathogenesis and drug response. However, to identify causal methylations out of a large number of potential polymorphic DNA methylation sites is challenging. This high-dimensional data brings two obstacles: first, many established statistical models are not scalable to so many features; second, multiple-test and overfitting become serious. To this end, a method to quickly filter candidate sites to narrow down targets for downstream analyses is urgently needed. BACkPAy is a pre-screening Bayesian approach to detect biological meaningful patterns of potential differential methylation levels with small sample size. BACkPAy prioritizes potentially important biomarkers by the Bayesian false discovery rate (FDR) approach. It filters non-informative sites (i.e., non-differential) with flat methylation pattern levels across experimental conditions. In this work, we applied BACkPAy to a genome-wide methylation dataset with three tissue types and each type contains three gastric cancer samples. We also applied LIMMA (Linear Models for Microarray and RNA-Seq Data) to compare its results with what we achieved by BACkPAy. Then, Cox proportional hazards regression models were utilized to visualize prognostics significant markers with The Cancer Genome Atlas (TCGA) data for survival analysis. Using BACkPAy, we identified eight biological meaningful patterns/groups of differential probes from the DNA methylation dataset. Using TCGA data, we also identified five prognostic genes (i.e., predictive to the progression of gastric cancer) that contain some differential methylation probes, whereas no significant results was identified using the Benjamin-Hochberg FDR in LIMMA. We showed the importance of using BACkPAy for the analysis of DNA methylation data with extremely small sample size in gastric cancer. We revealed that RDH13, CLDN11, TMTC1, UCHL1, and FOXP2 can serve as predictive biomarkers for gastric cancer treatment and the promoter methylation level of these five genes in serum could have prognostic and diagnostic functions in gastric cancer patients.

Maternal trauma and fear history predict BDNF methylation and gene expression in newborns

Trauma and related fear exert significant influence on mental and physical health throughout the lifespan and are associated with intergenerational patterns of development, health, and behavior. DNA methylation and gene expression are involved in our developmental adaptations to our experiences and can be influenced by social interventions. Patterns of DNA methylation and expression of a gene involved in neurodevelopment and psychiatric risk (BDNF) have been linked with childhood trauma. Given the intergenerational patterns of health and behavior, and previous links between childhood trauma and BDNF methylation and expression, this study investigated the potential for maternal history of traumatic experiences to influence development in her newborn, via changes in her newborn’s BDNF methylation and expression. We found that mothers’ trauma history was associated with epigenetic regulation of BDNF in their newborns. Moreover, the association between maternal trauma and BDNF methylation and expression patterns were moderated by newborn sex. Male newborns showed increased BDNF expression with maternal exposure to child abuse (p = .001), and increased BDNF methylation with greater maternal fear (p = .001). Female newborns showed reduced BDNF expression with greater maternal fear (p = .004). Practitioners strive to identify prevention and intervention avenues that will reduce the harmful effects of trauma. Future research should consider the potential for maternal historical trauma experiences to influence offspring DNA methylation and gene expression in a manner that could alter development and inform novel prevention strategies.

Differential methylation values in differential methylation analysis

MOTIVATION

Both β-value and M-value have been used as metrics to measure methylation levels. The M-value is more statistically valid for the differential analysis of methylation levels. However, the β-value is much more biologically interpretable and needs to be reported when M-value method is used for conducting differential methylation analysis. There is an urgent need to know how to interpret the degree of differential methylation from the M-value. In M-value linear regression model, differential methylation M-value ΔM can be easily obtained from the coefficient estimate, but it is not straightforward to get the differential methylation β-value, Δβ since it cannot be obtained from the coefficient alone.

RESULTS

To fill the gap, we have built a bridge to connect the statistically sound M-value linear regression model and the biologically interpretable Δβ. In this article, three methods were proposed to calculate differential methylation values, Δβ from M-value linear regression model and compared with the Δβ directly obtained from β-value linear regression model. We showed that under the condition that M-value linear regression model is correct, the method M-model-coef is the best among the four methods. M-model-M-mean method works very well too. If the coefficients α0, α2,…αp are not given (as 'MethLAB' package), the M-model-M-mean method should be used. The Δβ directly obtained from β-value linear regression model can give very biased results, especially when M-values are not in (-2, 2) or β-values are not in (0.2, 0.8).

AVAILABILITY AND IMPLEMENTATION

The dataset for example is available at the National Center for Biotechnology Information Gene Expression Omnibus repository, GSE104778.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

DNA Methylation Changes in Whole Blood and CD16+ Neutrophils in Response to Chronic Folic Acid Supplementation in Women of Childbearing Age

Folate, a water-soluble vitamin, is a key source of one-carbon groups for DNA methylation, but studies of the DNA methylation response to supplemental folic acid yield inconsistent results. These studies are commonly conducted using whole blood, which contains a mixed population of white blood cells that have been shown to confound results. The objective of this study was to determine if CD16+ neutrophils may provide more specific data than whole blood for identifying DNA methylation response to chronic folic acid supplementation. The study was performed in normal weight (BMI 18.5 - 24.9 kg/m2) women (18 - 35 y; n = 12), with blood samples taken before and after 8 weeks of folic acid supplementation at 800 μg/day. DNA methylation patterns from whole blood and isolated CD16+ neutrophils were measured across >485,000 CpG sites throughout the genome using the Infinium HumanMethylation450 BeadChip. Over the course of the 8-week supplementation, 6746 and 7513 CpG sites changed (p < 0.05) in whole blood and CD16+ neutrophils, respectively. DNA methylation decreased in 68.4% (whole blood) and 71.8% (CD16+ neutrophils) of these sites. There were only 182 CpG sites that changed in both the whole blood and CD16+ neutrophils, 139 of which changed in the same direction. These results suggest that the genome-wide DNA methylation response to chronic folic acid supplementation is different between whole blood and CD16+ neutrophils and that a single white blood cell type may function as a more specific epigenetic reporter of folate status than whole blood.

CD4+ and CD8+ T-Cell-Specific DNA Cytosine Methylation Differences Associated With Obesity

OBJECTIVE

Lifestyle factors associated with obesity may alter epigenome-regulated gene expression. Most studies examining epigenetic changes in obesity have analyzed DNA 5´-methylcytosine (5mC) in whole blood, representing a weighted average of several distantly related and regulated leukocyte classes. To examine leukocyte-specific differences associated with obesity, a pilot study examining 5mC in three distinct leukocyte types isolated from peripheral blood of women with normal weight and obesity was conducted.

METHODS

CD4⁺ T cells, CD8⁺ T cells, and CD16⁺ neutrophils were reiteratively isolated from blood, and 5mC levels were measured across >450,000 CG sites.

RESULTS

Nineteen CG sites were differentially methylated between women with obesity and with normal weight in CD4⁺ cells, 16 CG sites in CD8⁺ cells, and 0 CG sites in CD16⁺ neutrophils (q < 0.05). There were no common differentially methylated sites between the T-cell types. The amount of visceral adipose tissue was strongly associated with the methylation level of 79 CG sites in CD4⁺ cells, including 4 CG sites in CLSTN1's promoter, which, this study shows, may regulate its expression.

CONCLUSIONS

The methylomes of various leukocytes respond differently to obesity and levels of visceral adipose tissue. Highly significant differentially methylated sites in CD4⁺ and CD8⁺ cells in women with obesity that have apparent biological relevance to obesity were identified.

Effects of maternal obesity on Wharton's Jelly mesenchymal stromal cells

We investigated whether maternal metabolic environment affects mesenchymal stromal/stem cells (MSCs) from umbilical cord’s Wharton’s Jelly (WJ) on a molecular level, and potentially render them unsuitable for clinical use in multiple recipients. In this pilot study on umbilical cords post partum from healthy non-obese (BMI = 19–25; n = 7) and obese (BMI ≥ 30; n = 7) donors undergoing elective Cesarean section, we found that WJ MSC from obese donors showed slower population doubling and a stronger immunosuppressive activity. Genome-wide DNA methylation of triple positive (CD73⁺CD90⁺CD105⁺) WJ MSCs found 67 genes with at least one CpG site where the methylation difference was ≥0.2 in four or more obese donors. Only one gene, PNPLA7, demonstrated significant difference on methylome, transcriptome and protein level. Although the number of analysed donors is limited, our data suggest that the altered metabolic environment related to excessive body weight might bear consequences on the WJ MSCs.

Bioinformatics Tools for Genome-Wide Epigenetic Research

Epigenetics play a central role in the regulation of many important cellular processes, and dysregulations at the epigenetic level could be the source of serious pathologies, such as neurological disorders affecting brain development, neurodegeneration, and intellectual disability. Despite significant technological advances for epigenetic profiling, there is still a need for a systematic understanding of how epigenetics shapes cellular circuitry, and disease pathogenesis. The development of accurate computational approaches for analyzing complex epigenetic profiles is essential for disentangling the mechanisms underlying cellular development, and the intricate interaction networks determining and sensing chromatin modifications and DNA methylation to control gene expression. In this chapter, we review the recent advances in the field of "computational epigenetics," including computational methods for processing different types of epigenetic data, prediction of chromatin states, and study of protein dynamics. We also discuss how "computational epigenetics" has complemented the fast growth in the generation of epigenetic data for uncovering the main differences and similarities at the epigenetic level between individuals and the mechanisms underlying disease onset and progression.

Methylation of the oxytocin receptor gene mediates the effect of adversity on negative schemas and depression

Building upon various lines of research, we posited that methylation of the oxytocin receptor gene (OXTR) would mediate the effect of adult adversity on increased commitment to negative schemas and in turn the development of depression. We tested our model using structural equation modeling and longitudinal data from a sample of 100 middle-aged, African American women. The results provided strong support for the model. Analysis of the 12 CpG sites available for the promoter region of the OXTR gene identified four factors. One of these factors was related to the study variables, whereas the others were not. This factor mediated the effect of adult adversity on schemas relating to pessimism and distrust, and these schemas, in turn, mediated the impact of OXTR methylation on depression. All indirect effects were statistically significant, and they remained significant after controlling for childhood trauma, age, romantic relationship status, individual differences in cell types, and average level of genome-wide methylation. These finding suggest that epigenetic regulation of the oxytocin system may be a mechanism whereby the negative cognitions central to depression become biologically embedded.

DNA methylation analysis of paediatric low-grade astrocytomas identifies a tumour-specific hypomethylation signature in pilocytic astrocytomas

Low-grade gliomas (LGGs) account for about a third of all brain tumours in children. We conducted a detailed study of DNA methylation and gene expression to improve our understanding of the biology of pilocytic and diffuse astrocytomas. Pilocytic astrocytomas were found to have a distinctive signature at 315 CpG sites, of which 312 were hypomethylated and 3 were hypermethylated. Genomic analysis revealed that 182 of these sites are within annotated enhancers. The signature was not present in diffuse astrocytomas, or in published profiles of other brain tumours and normal brain tissue. The AP-1 transcription factor was predicted to bind within 200 bp of a subset of the 315 differentially methylated CpG sites; the AP-1 factors, FOS and FOSL1 were found to be up-regulated in pilocytic astrocytomas. We also analysed splice variants of the AP-1 target gene, CCND1, which encodes cell cycle regulator cyclin D1. CCND1a was found to be highly expressed in both pilocytic and diffuse astrocytomas, but diffuse astrocytomas have far higher expression of the oncogenic variant, CCND1b. These findings highlight novel genetic and epigenetic differences between pilocytic and diffuse astrocytoma, in addition to well-described alterations involving BRAF, MYB and FGFR1.

ADMIRE: analysis and visualization of differential methylation in genomic regions using the Infinium HumanMethylation450 Assay

Background

DNA methylation at cytosine nucleotides constitutes epigenetic gene regulation impacting cellular development and a wide range of diseases. Cytosine bases of the DNA are converted to 5-methylcytosine by the methyltransferase enzyme, acting as a reversible regulator of gene expression. Due to its outstanding importance in the epigenetic field, a number of lab techniques were developed to interrogate DNA methylation on a global range. Besides whole-genome bisulfite sequencing, the Infinium HumanMethylation450 Assay represents a versatile and cost-effective tool to investigate genome-wide changes of methylation patterns.

Results

Analysis of DNA Methylation In genomic REgions (ADMIRE) is an open source, semi-automatic analysis pipeline and visualization tool for Infinium HumanMethylation450 Assays with a special focus on ease of use. It features flexible experimental settings, quality control, automatic filtering, normalization, multiple testing, and differential analyses on arbitrary genomic regions. Publication-ready graphics, genome browser tracks, and table outputs include summary data and statistics, permitting instant comparison of methylation profiles between sample groups and the exploration of methylation patterns along the whole genome. ADMIREs statistical approach permits simultaneous large-scale analyses of hundreds of assays with little impact on algorithm runtimes.

Conclusions

The web-based version of ADMIRE provides a simple interface to researchers with limited programming skills, whereas the offline version is suitable for integration into custom pipelines. ADMIRE may be used via our freely available web service at https://bioinformatics.mpi-bn.mpg.de without any limitations concerning the size of a project. An offline version for local execution is available from our website or GitHub (https://github.molgen.mpg.de/loosolab/admire).

Electronic supplementary material

The online version of this article (doi:10.1186/s13072-015-0045-1) contains supplementary material, which is available to authorized users.

DNA methylation provides insight into intergenerational risk for preterm birth in African Americans

African Americans are at increased risk for spontaneous preterm birth (PTB). Though PTB is heritable, genetic studies have not identified variants that account for its intergenerational risk, prompting the hypothesis that epigenetic factors may also contribute. The objective of this study was to evaluate DNA methylation from maternal leukocytes to identify patterns specific to PTB and its intergenerational risk. DNA from peripheral leukocytes from African American women that delivered preterm (24-34 weeks; N = 16) or at term (39-41 weeks; N = 24) was assessed for DNA methylation using the HumanMethylation450 BeadChip. In maternal samples, 17,829 CpG sites associated with PTB, but no CpG site remained associated after correction for multiple comparisons. Examination of paired maternal-fetal samples identified 5,171 CpG sites in which methylation of maternal samples correlated with methylation of her respective fetus (FDR < 0.05). These correlated sites were enriched for association with PTB in maternal leukocytes. The majority of correlated CpG sites could be attributed to one or more genetic variants. They were also significantly more likely to be in genes involved in metabolic, cardiovascular, and immune pathways, suggesting a role for genetic and environmental contributions to PTB risk and chronic disease. The results of this study may provide insight into the factors underlying intergenerational risk for PTB and its consequences.

Higher levels of protective parenting are associated with better young adult health: exploration of mediation through epigenetic influences on pro-inflammatory processes

The current investigation was designed to examine the association of parenting during late childhood and early adolescence, a time of rapid physical development, with biological propensity for inflammation. Based on life course theory, it was hypothesized that parenting during this period of rapid growth and development would be associated with biological outcomes and self-reported health assessed in young adulthood. It was expected that association of parenting with health would be mediated either by effects on methylation of a key inflammatory factor, Tumor necrosis factor (TNF), or else by association with a pro-inflammatory shift in the distribution of mononuclear blood cells. Supporting expectations, in a sample of 398 African American youth residing in rural Georgia, followed from age 11 to age 19, parenting at ages 11-13 was associated with youth reports of better health at age 19. We found that parenting was associated with changes in TNF methylation as well as with changes in cell-type composition. However, whereas methylation of TNF was a significant mediator of the association of parenting with young adult health, variation in mononuclear white blood cell types was not a significant mediator of the association of parenting with young adult health. The current research suggests the potential value of examining the health-related effects of parenting in late childhood and early adolescence. Further examination of protection against pro-inflammatory tendencies conferred by parenting appears warranted.

Maternal adiposity negatively influences infant brain white matter development

OBJECTIVE

To study potential effects of maternal body composition on central nervous system (CNS) development of newborn infants.

METHODS

Diffusion tensor imaging (DTI) was used to evaluate brain white matter development in 2-week-old, full-term, appropriate for gestational age (AGA) infants from uncomplicated pregnancies of normal-weight (BMI < 25 at conception) or obese ( BMI = 30 at conception) and otherwise healthy mothers. Tract-based spatial statistics (TBSS) analyses were used for voxel-wise group comparison of fractional anisotropy (FA), a sensitive measure of white matter integrity. DNA methylation analyses of umbilical cord tissue focused on genes known to be important in CNS development were also performed.

RESULTS

Newborns from obese women had significantly lower FA values in multiple white matter regions than those born of normal-weight mothers. Global and regional FA values negatively correlated (P < 0.05) with maternal fat mass percentage. Linear regression analysis followed by gene ontology enrichment showed that methylation status of 68 CpG sites representing 57 genes with GO terms related to CNS development was significantly associated with maternal adiposity status.

CONCLUSIONS

These results suggest a negative association between maternal adiposity and white matter development in offspring.

Gene-specific differential DNA methylation and chronic arsenic exposure in an epigenome-wide association study of adults in Bangladesh

BACKGROUND

Inorganic arsenic is one of the most common naturally occurring contaminants found in the environment. Arsenic is associated with a number of health outcomes, with epigenetic modification suggested as a potential mechanism of toxicity.

OBJECTIVE

Among a sample of 400 adult participants, we evaluated the association between arsenic exposure, as measured by blood and urinary total arsenic concentrations, and epigenome-wide white blood cell DNA methylation.

METHODS

We used linear regression models to examine the associations between arsenic exposure and methylation at each CpG site, adjusted for sex, age, and batch. Differentially methylated loci were subsequently examined in relation to corresponding gene expression for functional evidence of gene regulation.

RESULTS

In adjusted analyses, we observed four differentially methylated CpG sites with urinary total arsenic concentration and three differentially methylated CpG sites with blood arsenic concentration, based on the Bonferroni-corrected significance threshold of p < 1 × 10(-7). Methylation of PLA2G2C (probe cg04605617) was the most significantly associated locus in relation to both urinary (p = 3.40 × 10(-11)) and blood arsenic concentrations (p = 1.48 × 10(-11)). Three additional novel methylation loci-SQSTM1 (cg01225779), SLC4A4 (cg06121226), and IGH (cg13651690)--were also significantly associated with arsenic exposure. Further, there was evidence of methylation-related gene regulation based on gene expression for a subset of differentially methylated loci.

CONCLUSIONS

We observed significant associations between arsenic exposure and gene-specific differential white blood cell DNA methylation, suggesting that epigenetic modifications may be an important pathway underlying arsenic toxicity. The specific differentially methylated loci identified may inform potential pathways for future interventions.

Ethnicity and Smoking-Associated DNA Methylation Changes at HIV Co-Receptor GPR15

Smoking is associated with poorer health outcomes for both African and European Americans. In order to better understand whether ethnic-specific genetic variation may underlie some of these differences, we compared the smoking-associated genome-wide methylation signatures of African Americans with those of European Americans, and followed up this analysis with a focused examination of the most ethnically divergent locus, cg19859270, at the GPR15 gene. We examined the association of methylation at this locus to the rs2230344 SNP and GPR15 gene and protein expression. Consistent with prior analyses, AHRR residue cg05575921 was the most differentially methylated residue in both African Americans and European Americans. However, the second most differentially methylated locus in African Americans, cg19859270, was only modestly differentially methylated in European Americans. Interrogation of the methylation status of this CpG residue found in GPR15, a chemokine receptor involved in HIV pathogenesis, showed a significant interaction of ethnicity with smoking as well as a marginal effect of genotype at rs2230344, a neighboring non-synonymous SNP, but only among African Americans. Gene and protein expression analyses showed that demethylation at cg19859270 was associated with an increase in both mRNA and protein levels. Since GPR15 is involved in the early stages of viral replication for some HIV-1 and HIV-2 isolates, and the prevalence of HIV is increased in African Americans and smokers, these data support a possible role for GPR15 in the ethnically dependent differential prevalence of HIV.

Comprehensive analysis of DNA methylation data with RnBeads

RnBeads is a software tool for large-scale analysis and interpretation of DNA methylation data, providing a user-friendly analysis workflow that yields detailed hypertext reports (http://rnbeads.mpi-inf.mpg.de/). Supported assays include whole-genome bisulfite sequencing, reduced representation bisulfite sequencing, Infinium microarrays and any other protocol that produces high-resolution DNA methylation data. Notable applications of RnBeads include the analysis of epigenome-wide association studies and epigenetic biomarker discovery in cancer cohorts.

Aberrant methylation of ERBB pathway genes in sporadic colorectal cancer

The ErbB signalling network plays a crucial role in the growth and progression of several cancers, including colorectal cancer (CRC), and includes potentially drug-targetable genes. Oncogenic activation of the ErbB pathway by mutations and focal amplifications have emerged recently as an important predictive marker of the prognosis of CRC patients. However, in contrast to genetic events, little is known about epigenetic alternations of ErbB-associated genes and their impact on gene expression. Genome-wide methylation in sporadic CRCs (n = 12) paired with adjacent normal tissues have been previously analysed by Illumina Infinium HumanMethylation27 (HM27) at 27,578 CpG sites. For confirmation of our initial genome-wide analysis, we used a published HM27 dataset (GSE25062). Subsequently, CpG island methylation of selected ErbB pathway-associated genes was assessed on 233 CRC samples using methylation-sensitive polymerase chain reaction (MS-PCR) and analysed along with various genetic factors associated with CRC [epigenotype, BRAF and KRAS mutations, microsatellite instability (MSI)]. Methylation and expression integration was performed using published datasets including 25 pairs of CRC and normal colon tissues (GSE25062 and GSE25070), and confirmed with real-time PCR. Our previous microarray-based genome-wide DNA methylation analysis of 12 CRCs revealed that four ErbB-associated genes (PIK3CD, PKCΒ, ERBB4, ) were differentially methylated in CRCs. This was further confirmed by statistical re-analysis of an HM27 dataset (GSE25062). Frequent methylation at these loci in tumours was subsequently confirmed by MS-PCR (63 %, 43 %, 43 % and 92 %, respectively). Hypermethylation of PKCΒ associated with KRAS mutation (p = 0.04), whereas hypermethylation of ERBB4 associated with high-methylation epigenotypes (HME), BRAF mutation and MSI (p = 0.001, 0.002 and 0.0002, respectively). One of the four analysed genes (PKCΒ) was significantly downregulated in CRC tissue, as revealed by real-time PCR and re-analysis of the GSE25062 and GSE25070 datasets. After careful re-analysis of published methylation and expression data, we conclude that methylation of ERBB4, PAK7 and PIK3CD has no functional role in CRC carcinogenesis. In contrast, methylation seems to have a potential impact on the biology of colorectal tumours by negatively modulating the expression of PKCΒ. Importantly, the relationship between DNA methylation of PKCΒ and gene expression may warrant further attention in the context of colon cancer chemoprevention and anti-cancer therapy.

The CpG island methylator phenotype in breast cancer is associated with the lobular subtype

BACKGROUND

Aberrations in DNA methylation patterns are well-described in human malignancies. However, the existence of the 'CpG island methylator phenotype' (CIMP) in human breast cancer is still controversial.

MATERIALS & METHODS

Illumina's HumanMethylation 450K BeadChip was used to analyze genome-wide DNA methylation patterns. Chromosomal abnormalities were determined by array-based CGH.

RESULTS

Invasive lobular breast carcinomas exhibit the highest number of differentially methylated CpG sites and a strong inverse correlation of aberrant DNA hypermethylation and copy number alterations. Nine differentially methylated regions within seven genes discriminating the investigated subgroups were identified and validated in an independent validation cohort and correlated to a better relapse-free survival.

CONCLUSION

These results depict a clear difference between genetically and epigenetically unstable breast carcinomas indicating different ways of tumor progression and/or initiation, which strongly supports the association of CIMP with the lobular subtype and provide new options for detection and therapy.

A pilot examination of the genome-wide DNA methylation signatures of subjects entering and exiting short-term alcohol dependence treatment programs

Alcoholism has a profound impact on millions of people throughout the world. However, the ability to determine if a patient needs treatment is hindered by reliance on self-reporting and the clinician's capability to monitor the patient's response to treatment is challenged by the lack of reliable biomarkers. Using a genome-wide approach, we have previously shown that chronic alcohol use is associated with methylation changes in DNA from human cell lines. In this pilot study, we now examine DNA methylation in peripheral mononuclear cell DNA gathered from subjects as they enter and leave short-term alcohol treatment. When compared with abstinent controls, subjects with heavy alcohol use show widespread changes in DNA methylation that have a tendency to reverse with abstinence. Pathway analysis demonstrates that these changes map to gene networks involved in apoptosis. There is no significant overlap of the alcohol signature with the methylation signature previously derived for smoking. We conclude that DNA methylation may have future clinical utility in assessing acute alcohol use status and monitoring treatment response.

Comparisons of non-Gaussian statistical models in DNA methylation analysis

As a key regulatory mechanism of gene expression, DNA methylation patterns are widely altered in many complex genetic diseases, including cancer. DNA methylation is naturally quantified by bounded support data; therefore, it is non-Gaussian distributed. In order to capture such properties, we introduce some non-Gaussian statistical models to perform dimension reduction on DNA methylation data. Afterwards, non-Gaussian statistical model-based unsupervised clustering strategies are applied to cluster the data. Comparisons and analysis of different dimension reduction strategies and unsupervised clustering methods are presented. Experimental results show that the non-Gaussian statistical model-based methods are superior to the conventional Gaussian distribution-based method. They are meaningful tools for DNA methylation analysis. Moreover, among several non-Gaussian methods, the one that captures the bounded nature of DNA methylation data reveals the best clustering performance.

Epigenetic changes associated with inflammation in breast cancer patients treated with chemotherapy

Inflammation has been associated with fatigue during and after various types of breast cancer treatments. We examined whether prior chemotherapy was associated with DNA methylation patterns that could explain persisting inflammation and/or fatigue in women treated for breast cancer. Prior to breast radiation therapy, DNA was extracted from peripheral blood mononuclear cells (PBMCs) of 61 Stage 0-IIIA breast cancer patients who had received partial mastectomy with or without chemotherapy. DNA methylation was assessed at >485,000 CpG sites across the genome along with fatigue and plasma inflammatory markers previously associated with fatigue. Compared to non-chemotherapy-treated, women who had received chemotherapy exhibited significantly decreased methylation at eight CpG sites (p<1.03×10(-7)) including four in exon 11 of transmembrane protein 49 (TMEM49), which demonstrated the largest decreases in methylation. Lower methylation at each identified CpG site was associated with increased plasma soluble tumor necrosis factor receptor 2 (sTNFR2) and interleukin (IL)-6 and mediated the relationship between chemotherapy and increases in these inflammatory biomarkers adjusting for multiple clinical and treatment characteristics. sTNFR2, but not CpG methylation status, was correlated with fatigue. Six months after breast radiation therapy, DNA methylation, inflammatory biomarkers and fatigue assessments were repeated in a subset of subjects (N=39). Reduced methylation in 4 of the 8 identified CpG sites was still observed in chemotherapy versus non-chemotherapy-treated patients, albeit with some decay indicating the dynamic and potentially reversible nature of the changes. Reduced methylation in these 4 CpG sites also continued to correlate with either increased sTNFR2 or IL-6, but not fatigue. In conclusion, prior chemotherapy treatment was associated with decreased methylation of CpG sites in DNA from PBMCs of breast cancer patients, which correlated with increased inflammatory markers prior to and 6months after radiation therapy. Persisting epigenetic changes secondary to chemotherapy may be one factor that contributes to inflammation and its consequences including cancer-related fatigue in vulnerable breast cancer patients.

The effect of smoking on DNA methylation of peripheral blood mononuclear cells from African American women

Background

Regular smoking is associated with a wide variety of syndromes with prominent inflammatory components such as cancer, obesity and type 2 diabetes. Heavy regular smoking is also associated with changes in the DNA methylation of peripheral mononuclear cells. However, in younger smokers, inflammatory epigenetic findings are largely absent which suggests the inflammatory response(s) to smoking may be dose dependent. To help understand whether peripheral mononuclear cells have a role in mediating these responses in older smokers with higher cumulative smoke exposure, we examined genome-wide DNA methylation in a group of well characterized adult African American subjects informative for smoking, as well as serum C-reactive protein (CRP) and interleukin-6 receptor (IL6R) levels. In addition, complementary bioinformatic analyses were conducted to delineate possible pathways affected by long-term smoking.

Results

Genome-wide DNA methylation analysis with respect to smoking status yielded 910 significant loci after Benjamini-Hochberg correction. In particular, two loci from the AHRR gene (cg05575921 and cg23576855) and one locus from the GPR15 gene (cg19859270) were identified as highly significantly differentially methylated between smokers and non-smokers. The bioinformatic analyses showed that long-term chronic smoking is associated with altered promoter DNA methylation of genes coding for proteins mapping to critical sub-networks moderating inflammation, immune function, and coagulation.

Conclusions

We conclude that chronic regular smoking is associated with changes in peripheral mononuclear cell methylation signature which perturb inflammatory and immune function pathways and may contribute to increased vulnerability for complex illnesses with inflammatory components.

Changes in DNA methylation at the aryl hydrocarbon receptor repressor may be a new biomarker for smoking

Background

Smoking is the largest preventable cause of morbidity and mortality in the United States. In previous work, we demonstrated that altered DNA methylation at the aryl hydrocarbon receptor repressor (AHRR) is correlated with self-reported smoking in 19-year-old African Americans with relatively low levels of smoking. However, one limitation of the prior work is that it was based on self-reported data only. Therefore, the relationship of AHRR methylation to smoking in older subjects and to indicators such as serum cotinine levels remains unknown. To address this question, we examined the relationship between genome- wide DNA methylation and smoking status as indicated by serum cotinine levels in a cohort of 22-year-old African American men.

Results

Consistent with prior findings, smoking was associated with significant DNA demethylation at two distinct loci within AHRR (cg05575921 and cg21161138) with the degree of demethylation being greater than that observed in the prior cohort of 19-year-old smoking subjects. Additionally, methylation status at the AHRR residue interrogated by cg05575921 was highly correlated with serum cotinine levels (adjusted R² = 0.42, P < 0.0001).

Conclusions

We conclude that AHRR DNA methylation status is a sensitive marker of smoking history and could serve as a biomarker of smoking that could supplement self-report or existing biomarker measures in clinical or epidemiological analyses of the effects of smoking. In addition, if properly configured as a clinical assay, the determination of AHRR methylation could also be used as a screening tool in efforts to target antismoking interventions to nascent smokers in the early phases of smoking.

Review of processing and analysis methods for DNA methylation array data

The promise of epigenome-wide association studies and cancer-specific somatic DNA methylation changes in improving our understanding of cancer, coupled with the decreasing cost and increasing coverage of DNA methylation microarrays, has brought about a surge in the use of these technologies. Here, we aim to provide both a review of issues encountered in the processing and analysis of array-based DNA methylation data and a summary of the advantages of recent approaches proposed for handling those issues, focusing on approaches publicly available in open-source environments such as R and Bioconductor. We hope that the processing tools and analysis flowchart described herein will facilitate researchers to effectively use these powerful DNA methylation array-based platforms, thereby advancing our understanding of human health and disease.

Epigenomic association analysis identifies smoking-related DNA methylation sites in African Americans

Cigarette smoking is an environmental risk factor for many chronic diseases, and disease risk can often be managed by smoking control. Smoking can induce cellular and molecular changes, including epigenetic modification, but the short- and long-term epigenetic modifications caused by cigarette smoking at the gene level have not been well understood. Recent studies have identified smoking-related DNA methylation (DNAm) sites in Caucasians. To determine whether the same DNAm sites associate with smoking in African Americans, and to identify novel smoking-related DNAm sites, we conducted a methylome-wide association study of cigarette smoking using a discovery sample of 972 African Americans, and a replication sample of 239 African Americans with two array-based methods. Among 15 DNAm sites significantly associated with smoking after correction for multiple testing in our discovery sample, 5 DNAm sites are replicated in an independent cohort, and 14 sites in the replication sample have effects in the same direction as in the discovery sample. The top two smoking-related DNAm sites in F2RL3 (factor II receptor-like 3) and GPR15 (G-protein-coupled receptor 15) observed in African Americans are consistent with previous findings in Caucasians. The associations between the replicated DNAm sites and smoking remain significant after adjusting for genetic background. Despite the distinct genetic background between African Americans and Caucasians, the DNAm from the two ethnic groups shares common associations with cigarette smoking, which suggests a common molecular mechanism of epigenetic modification influenced by environmental exposure.

Fetal DNA Methylation Associates with Early Spontaneous Preterm Birth and Gestational Age

Spontaneous preterm birth (PTB, <37 weeks gestation) is a major public health concern, and children born preterm have a higher risk of morbidity and mortality throughout their lives. Recent studies suggest that fetal DNA methylation of several genes varies across a range of gestational ages (GA), but it is not yet clear if fetal epigenetic changes associate with PTB. The objective of this study is to interrogate methylation patterns across the genome in fetal leukocyte DNA from African Americans with early PTB (24^1/7–34^0/7 weeks; N = 22) or term births (39^0/7–40^6/7weeks; N = 28) and to evaluate the association of each CpG site with PTB and GA. DNA methylation was assessed across the genome with the HumanMethylation450 BeadChip. For each individual sample and CpG site, the proportion of DNA methylation was estimated. The associations between methylation and PTB or GA were evaluated by fitting a separate linear model for each CpG site, adjusting for relevant covariates. Overall, 29 CpG sites associated with PTB (FDR<.05; 5.7×10⁻¹⁰<p<2.9×10⁻⁶) independent of GA. Also, 9637 sites associated with GA (FDR<.05; 9.5×10⁻¹⁶<p<1.0×10⁻³), with 61.8% decreasing in methylation with shorter GA. GA-associated CpG sites were depleted in the CpG islands of their respective genes (p<2.2×10⁻¹⁶). Gene set enrichment analysis (GSEA) supported enrichment of GA-associated CpG sites in genes that play a role in embryonic development as well as the extracellular matrix. Additionally, this study replicated the association of several CpG sites associated with gestational age in other studies (CRHBP, PIK3CD and AVP). Dramatic differences in fetal DNA methylation are evident in fetuses born preterm versus at term, and the patterns established at birth may provide insight into the long-term consequences associated with PTB.

COHCAP: an integrative genomic pipeline for single-nucleotide resolution DNA methylation analysis

COHCAP (City of Hope CpG Island Analysis Pipeline) is an algorithm to analyze single-nucleotide resolution DNA methylation data produced by either an Illumina methylation array or targeted bisulfite sequencing. The goal of the COHCAP algorithm is to identify CpG islands that show a consistent pattern of methylation among CpG sites. COHCAP is currently the only DNA methylation package that provides integration with gene expression data to identify a subset of CpG islands that are most likely to regulate downstream gene expression, and it can generate lists of differentially methylated CpG islands with ∼50% concordance with gene expression from both cell line data and heterogeneous patient data. For example, this article describes known breast cancer biomarkers (such as estrogen receptor) with a negative correlation between DNA methylation and gene expression. COHCAP also provides visualization for quality control metrics, regions of differential methylation and correlation between methylation and gene expression. This software is freely available at https://sourceforge.net/projects/cohcap/.

A variational Bayes beta mixture model for feature selection in DNA methylation studies

An increasing number of studies are using beadarrays to measure DNA methylation on a genome-wide basis. The purpose is to identify novel biomarkers in a wide range of complex genetic diseases including cancer. A common difficulty encountered in these studies is distinguishing true biomarkers from false positives. While statistical methods aimed at improving the feature selection step have been developed for gene expression, relatively few methods have been adapted to DNA methylation data, which is naturally beta-distributed. Here we explore and propose an innovative application of a recently developed variational Bayesian beta-mixture model (VBBMM) to the feature selection problem in the context of DNA methylation data generated from a highly popular beadarray technology. We demonstrate that VBBMM offers significant improvements in inference and feature selection in this type of data compared to an Expectation-Maximization (EM) algorithm, at a significantly reduced computational cost. We further demonstrate the added value of VBBMM as a feature selection and prioritization step in the context of identifying prognostic markers in breast cancer. A variational Bayesian approach to feature selection of DNA methylation profiles should thus be of value to any study undergoing large-scale DNA methylation profiling in search of novel biomarkers.

Analysing and interpreting DNA methylation data

DNA methylation is an epigenetic mark that has suspected regulatory roles in a broad range of biological processes and diseases. The technology is now available for studying DNA methylation genome-wide, at a high resolution and in a large number of samples. This Review discusses relevant concepts, computational methods and software tools for analysing and interpreting DNA methylation data. It focuses not only on the bioinformatic challenges of large epigenome-mapping projects and epigenome-wide association studies but also highlights software tools that make genome-wide DNA methylation mapping more accessible for laboratories with limited bioinformatics experience.

Demethylation of the aryl hydrocarbon receptor repressor as a biomarker for nascent smokers

Epigenetic modifications to peripheral white blood cell DNA occur in response to a wide variety of exposures. In prior work, we and others have shown that broad changes in DNA methylation, particularly at the aryl hydrocarbon receptor repressor (AHRR) locus, occur in samples from subjects with long histories of smoking. However, given the large number of epigenetic changes that occur in response to prolonged smoking, the primacy of the response at AHRR and the sensitivity of these changes to low levels of smoking are not known. Therefore, we examined the association of smoking to genome lymphocyte DNA methylation status in a representative sample of 399 African American youths living in the rural South that includes 72 subjects with less than one half-pack year of exposure. Consistent with our prior findings, we found a stepwise effect of smoking on DNA methylation among youth with relatively brief exposure histories at a CpG residue in AHRR (cg05575921) (FDR corrected p values; 3 × 10 (-7) and 0.09 in the male and female samples, respectively) that was identified in previous studies and at which the effects of smoking were significant, even in those subjects with less than one half pack year exposure. We conclude that AHRR demethylation at cg05575921 in peripheral cells may serve as an early, sensitive biomarker for even low levels of exposure to tobacco smoke, providing a non-self-report alternative for nascent exposure to tobacco smoke. We also suggest that the AHRR/AHR pathway may be functional in the response of peripheral white blood cells to tobacco smoke exposure.

Identification of DNA methylation biomarkers from Infinium arrays

Epigenetic modifications of DNA, such as cytosine methylation are differentially abundant in diseases such as cancer. A goal for clinical research is finding sites that are differentially methylated between groups of samples to act as potential biomarkers for disease outcome. However, clinical samples are often limited in availability, represent a heterogeneous collection of cells or are of uncertain clinical class. Array-based methods for identification of methylation provide a cost-effective method to survey a proportion of the methylome at single base resolution. The Illumina Infinium array has become a popular and reliable high throughput method in this field and are proving useful in the identification of biomarkers for disease. Here, we compare a commonly used statistical test with a new intuitive and flexible computational approach to quickly detect differentially methylated sites. The method rapidly identifies and ranks candidate lists with greatest inter-group variability whilst controlling for intra-group variability. Intuitive and biologically relevant filters can be imposed to quickly identify sites and genes of interest.