Estimation of blood cellular heterogeneity in newborns and children for epigenome-wide association studies

Decoding depression: a comprehensive multi-cohort exploration of blood DNA methylation using machine learning and deep learning approaches

The causes of depression are complex, and the current diagnosis methods rely solely on psychiatric evaluations with no incorporation of laboratory biomarkers in clinical practices. We investigated the stability of blood DNA methylation depression signatures in six different populations using six public and two domestic cohorts (n = 1942) conducting mega-analysis and meta-analysis of the individual studies. We evaluated 12 machine learning and deep learning strategies for depression classification both in cross-validation (CV) and in hold-out tests using merged data from 8 separate batches, constructing models with both biased and unbiased feature selection. We found 1987 CpG sites related to depression in both mega- and meta-analysis at the nominal level, and the associated genes were nominally related to axon guidance and immune pathways based on enrichment analysis and eQTM data. Random forest classifiers achieved the highest performance (AUC 0.73 and 0.76) in CV and hold-out tests respectively on the batch-level processed data. In contrast, the methylation showed low predictive power (all AUCs < 0.57) for all classifiers in CV and no predictive power in hold-out tests when used with harmonized data. All models achieved significantly better performance (>14% gain in AUCs) with pre-selected features (selection bias), with some of the models (joint autoencoder-classifier) reaching AUCs of up to 0.91 in the final testing regardless of data preparation. Different algorithmic feature selection approaches may outperform limma, however, random forest models perform well regardless of the strategy. The results provide an overview over potential future biomarkers for depression and highlight many important methodological aspects for DNA methylation-based depression profiling including the use of machine learning strategies.

Uncertainty quantification of reference-based cellular deconvolution algorithms

The majority of epigenetic epidemiology studies to date have generated genome-wide profiles from bulk tissues (e.g., whole blood) however these are vulnerable to confounding from variation in cellular composition. Proxies for cellular composition can be mathematically derived from the bulk tissue profiles using a deconvolution algorithm; however, there is no method to assess the validity of these estimates for a dataset where the true cellular proportions are unknown. In this study, we describe, validate and characterize a sample level accuracy metric for derived cellular heterogeneity variables. The CETYGO score captures the deviation between a sample's DNA methylation profile and its expected profile given the estimated cellular proportions and cell type reference profiles. We demonstrate that the CETYGO score consistently distinguishes inaccurate and incomplete deconvolutions when applied to reconstructed whole blood profiles. By applying our novel metric to >6,300 empirical whole blood profiles, we find that estimating accurate cellular composition is influenced by both technical and biological variation. In particular, we show that when using a common reference panel for whole blood, less accurate estimates are generated for females, neonates, older individuals and smokers. Our results highlight the utility of a metric to assess the accuracy of cellular deconvolution, and describe how it can enhance studies of DNA methylation that are reliant on statistical proxies for cellular heterogeneity. To facilitate incorporating our methodology into existing pipelines, we have made it freely available as an R package (https://github.com/ds420/CETYGO).

Investigating the effects of maltreatment and acute stress on the concordance of blood and DNA methylation methods of estimating immune cell proportions

BACKGROUND

Immune cell proportions can be used to detect pathophysiological states and are also critical covariates in genomic analyses. The complete blood count (CBC) is the most common method of immune cell proportion estimation, but immune cell proportions can also be estimated using whole-genome DNA methylation (DNAm). Although the concordance of CBC and DNAm estimations has been validated in various adult and clinical populations, less is known about the concordance of existing estimators among stress-exposed individuals. As early life adversity and acute psychosocial stress have both been associated with unique DNAm alterations, the concordance of CBC and DNAm immune cell proportion needs to be validated in various states of stress.

RESULTS

We report the correlation and concordance between CBC and DNAm estimates of immune cell proportions using the Illumina EPIC DNAm array within two unique studies: Study 1, a high-risk pediatric cohort of children oversampled for exposure to maltreatment (N = 365, age 8 to 14 years), and Study 2, a sample of young adults who have participated in an acute laboratory stressor with four pre- and post-stress measurements (N = 28, number of observations = 100). Comparing CBC and DNAm proportions across both studies, estimates of neutrophils (r = 0.948, p < 0.001), lymphocytes (r = 0.916, p < 0.001), and eosinophils (r = 0.933, p < 0.001) were highly correlated, while monocyte estimates were moderately correlated (r = 0.766, p < 0.001) and basophil estimates were weakly correlated (r = 0.189, p < 0.001). In Study 1, we observed significant deviations in raw values between the two approaches for some immune cell subtypes; however, the observed differences were not significantly predicted by exposure to child maltreatment. In Study 2, while significant changes in immune cell proportions were observed in response to acute psychosocial stress for both CBC and DNAm estimates, the observed changes were similar for both approaches.

CONCLUSIONS

Although significant differences in immune cell proportion estimates between CBC and DNAm exist, as well as stress-induced changes in immune cell proportions, neither child maltreatment nor acute psychosocial stress alters the concordance of CBC and DNAm estimation methods. These results suggest that the agreement between CBC and DNAm estimators of immune cell proportions is robust to exposure to child maltreatment and acute psychosocial stress.

Maternal exposure to cadmium during pregnancy is associated with changes in DNA methylation that are persistent at 9 years of age

BACKGROUND

Cadmium (Cd) exposure during gestation has been associated with altered DNA methylation at birth, but it is not known if the changes in methylation persist into childhood.

OBJECTIVES

To evaluate whether gestational Cd-related changes of DNA methylation persist from birth to 9 years of age.

METHODS

We studied mother-child dyads in a longitudinal cohort in rural Bangladesh. Cadmium concentrations in maternal blood (erythrocyte fraction; Ery-Cd) at gestational week 14 and in child urine (U-Cd, long-term exposure marker) at 9 years were measured using inductively coupled plasma mass spectrometry. The epigenome-wide DNA methylation was measured in mononuclear cells (PBMCs) prepared from cord blood and peripheral blood at 9 years in 71 children (hereafter referred to as the explorative group) by Infinium HumanMethylation450K BeadChip. Replication of one differentially methylated region (DMR; 9 CpG sites) was performed in PBMCs of 160 9-year-old children (validation group) by EpiTyper MALDI-TOF mass spectrometry.

RESULTS

The median maternal Ery-Cd concentration was 1.24 µg/kg (range 0.35, 4.55) in the explorative group and 0.83 µg/kg (0.08, 2.97) in the validation group. The median U-Cd concentration in the 9-year-old children was 0.26 µg/L (0.09, 1.06) in the explorative group and 0.32 µg/L (0.07, 1.33) in the validation group. In the explorative group, we identified ten DMRs, both in cord blood and in PBMCs at 9 years, that were associated with maternal Ery-Cd. Eight out of the ten DMRs were hypomethylated and three of the hypomethylated DMRs were located in the HLA region on chromosome 6. One of the DMRs (hypomethylated) in the HLA region (upstream of the zinc finger protein 57 homolog, ZFP57 gene) was replicated in the validation group, and we found that it was hypomethylated in relation to maternal Ery-Cd, but not child U-Cd.

CONCLUSION

Gestational exposure to Cd appears to be associated with regional changes, especially hypomethylated, in DNA methylation that linger from birth up to prepubertal age.

Enhanced cell deconvolution of peripheral blood using DNA methylation for high-resolution immune profiling

DNA methylation microarrays can be employed to interrogate cell-type composition in complex tissues. Here, we expand reference-based deconvolution of blood DNA methylation to include 12 leukocyte subtypes (neutrophils, eosinophils, basophils, monocytes, naïve and memory B cells, naïve and memory CD4 + and CD8 + T cells, natural killer, and T regulatory cells). Including derived variables, our method provides 56 immune profile variables. The IDOL (IDentifying Optimal Libraries) algorithm was used to identify libraries for deconvolution of DNA methylation data for current and previous platforms. The accuracy of deconvolution estimates obtained using our enhanced libraries was validated using artificial mixtures and whole-blood DNA methylation with known cellular composition from flow cytometry. We applied our libraries to deconvolve cancer, aging, and autoimmune disease datasets. In conclusion, these libraries enable a detailed representation of immune-cell profiles in blood using only DNA and facilitate a standardized, thorough investigation of immune profiles in human health and disease.

DeCompress: tissue compartment deconvolution of targeted mRNA expression panels using compressed sensing

Targeted mRNA expression panels, measuring up to 800 genes, are used in academic and clinical settings due to low cost and high sensitivity for archived samples. Most samples assayed on targeted panels originate from bulk tissue comprised of many cell types, and cell-type heterogeneity confounds biological signals. Reference-free methods are used when cell-type-specific expression references are unavailable, but limited feature spaces render implementation challenging in targeted panels. Here, we present DeCompress, a semi-reference-free deconvolution method for targeted panels. DeCompress leverages a reference RNA-seq or microarray dataset from similar tissue to expand the feature space of targeted panels using compressed sensing. Ensemble reference-free deconvolution is performed on this artificially expanded dataset to estimate cell-type proportions and gene signatures. In simulated mixtures, four public cell line mixtures, and a targeted panel (1199 samples; 406 genes) from the Carolina Breast Cancer Study, DeCompress recapitulates cell-type proportions with less error than reference-free methods and finds biologically relevant compartments. We integrate compartment estimates into cis-eQTL mapping in breast cancer, identifying a tumor-specific cis-eQTL for CCR3 (C-C Motif Chemokine Receptor 3) at a risk locus. DeCompress improves upon reference-free methods without requiring expression profiles from pure cell populations, with applications in genomic analyses and clinical settings.

Robust partial reference-free cell composition estimation from tissue expression

MOTIVATION

In the analysis of high-throughput omics data from tissue samples, estimating and accounting for cell composition have been recognized as important steps. High cost, intensive labor requirements and technical limitations hinder the cell composition quantification using cell-sorting or single-cell technologies. Computational methods for cell composition estimation are available, but they are either limited by the availability of a reference panel or suffer from low accuracy.

RESULTS

We introduce TOols for the Analysis of heterogeneouS Tissues TOAST/-P and TOAST/+P, two partial reference-free algorithms for estimating cell composition of heterogeneous tissues based on their gene expression profiles. TOAST/-P and TOAST/+P incorporate additional biological information, including cell-type-specific markers and prior knowledge of compositions, in the estimation procedure. Extensive simulation studies and real data analyses demonstrate that the proposed methods provide more accurate and robust cell composition estimation than existing methods.

AVAILABILITY AND IMPLEMENTATION

The proposed methods TOAST/-P and TOAST/+P are implemented as part of the R/Bioconductor package TOAST at https://bioconductor.org/packages/TOAST.

CONTACT

ziyi.li@emory.edu or hao.wu@emory.edu.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

5-Hydroxymethylcytosine in cord blood and associations of DNA methylation with sex in newborns

DNA methylation has been widely studied for associations with exposures and health outcomes. Both 5-methylcytosine (5mC) and 5-hydroxymethylcytosine (5hmC) are epigenetic marks that may function differently to impact gene expression; however, the most commonly used technology to assess methylation for population studies in blood use are the Illumina 450K and EPIC BeadChips, for which the traditional bisulfite conversion does not differentiate 5mC and 5hmC marks. We used a modified protocol originally developed by Stewart et al. to analyse oxidative bisulfite-converted and conventional bisulfite-converted DNA for the same subject in parallel by the EPIC chip, allowing us to isolate the two measures. We measured 5mC and 5hmC in cord blood of 41 newborn participants of the Center for Health Assessment of Mothers and Children of Salinas (CHAMACOS) birth cohort and investigated differential methylation of 5mC + 5hmC, isolated 5mC and isolated 5hmC with sex at birth as an example of a biological variable previously associated with DNA methylation. Results showed low levels of 5hmC throughout the epigenome in the cord blood samples in comparison to 5mC. The concordance of autosomal hits between 5mC + 5hmC and exclusive 5mC analyses were low (25%); however, overlap was larger with increased effect size difference. There were 43 autosomal cytosine nucleotide followed by a guanine nucleotide (CpG) sites where 5hmC was associated with sex, 21 of which were unique to 5hmC after adjustment for cell composition. 5hmC only accounts for a small portion of overall methylation in cord blood; however, it has the potential to impact interpretation of combined 5hmC + 5mC studies in cord blood, especially given that effect sizes of differential methylation analyses are often small. Several significant CpG sites were unique to 5hmC, suggesting some functions distinct from 5mC. More studies of genome-wide 5hmC in children are warranted.

TOAST: improving reference-free cell composition estimation by cross-cell type differential analysis

In the analysis of high-throughput data from complex samples, cell composition is an important factor that needs to be accounted for. Except for a limited number of tissues with known pure cell type profiles, a majority of genomics and epigenetics data relies on the "reference-free deconvolution" methods to estimate cell composition. We develop a novel computational method to improve reference-free deconvolution, which iteratively searches for cell type-specific features and performs composition estimation. Simulation studies and applications to six real datasets including both DNA methylation and gene expression data demonstrate favorable performance of the proposed method. TOAST is available at https://bioconductor.org/packages/TOAST .

Distinct epigenetic profiles in children with perinatally-acquired HIV on antiretroviral therapy

Perinatally-acquired HIV has persistent effects on long-term health outcomes, even after early treatment. We hypothesize that epigenetic indicators, such as DNA methylation, may elucidate cellular processes that explain these effects. Here, we compared DNA methylation profiles in whole blood from 120 HIV-infected children on antiretroviral therapy (ART) and 60 frequency age-matched HIV-uninfected children aged 4–9 years in Johannesburg, South Africa. Using an individual CpG site approach, we found 1,309 differentially-methylated (DM) CpG sites between groups, including 1,271 CpG sites that were hyper-methylated in the HIV-infected group and 38 CpG sites that were hypo-methylated in the HIV-infected group. Six hyper-methylated CpG sites were in EBF4, which codes for a transcription factor involved in B-cell maturation. The top hypomethylated site was in the promoter region of NLRC5, encoding a transcription factor that regulates major histocompatibility complex (MHC) class I molecule expression. Using a differentially-methylated region (DMR) approach, we found 315 DMRs between groups, including 28 regions encompassing 686 CpG sites on chromosome 6. A large number of the genes identified in both the CpG site and DMR approaches were located in the MHC region on chromosome 6, which plays an important role in the adaptive immune system. This study provides the first evidence that changes in the epigenome are detectable in children with perinatally-acquired HIV infection on suppressive ART started at an early age.

Pregnancy lipidomic profiles and DNA methylation in newborns from the CHAMACOS cohort

Lipids play a role in many biological functions and the newly emerging field of lipidomics aims to characterize the varying classes of lipid molecules present in biological specimens. Animal models have shown associations between maternal dietary supplementation with fatty acids during pregnancy and epigenetic changes in their offspring, demonstrating a mechanism through which prenatal environment can affect outcomes in children; however, data on maternal lipid metabolite levels during pregnancy and newborn DNA methylation in humans are sparse. In this study, we assessed the relationship of maternal lipid metabolites measured in the blood from pregnant women with newborn DNA methylation profiles in the Center for the Health Assessment of Mothers and Children of Salinas cohort. Targeted metabolomics was performed by selected reaction monitoring liquid chromatography and triple quadrupole mass spectrometry to measure 92 metabolites in plasma samples of pregnant women at ∼26 weeks gestation. DNA methylation was assessed using the Infinium HumanMethylation 450K BeadChip adjusting for cord blood cell composition. We uncovered numerous false discovery rate significant associations between maternal metabolite levels, particularly phospholipid and lysolipid metabolites, and newborn methylation. The majority of the observed relationships were negative, suggesting that higher lipid metabolites during pregnancy are associated with lower methylation levels at genes related to fetal development. These results further elucidate the complex relationship between early life exposures, maternal lipid metabolites, and infant epigenetic status.

BayesCCE: a Bayesian framework for estimating cell-type composition from DNA methylation without the need for methylation reference

We introduce a Bayesian semi-supervised method for estimating cell counts from DNA methylation by leveraging an easily obtainable prior knowledge on the cell-type composition distribution of the studied tissue. We show mathematically and empirically that alternative methods which attempt to infer cell counts without methylation reference only capture linear combinations of cell counts rather than provide one component per cell type. Our approach allows the construction of components such that each component corresponds to a single cell type, and provides a new opportunity to investigate cell compositions in genomic studies of tissues for which it was not possible before.

Prenatal arsenic exposure is associated with increased plasma IGFBP3 concentrations in 9-year-old children partly via changes in DNA methylation

Exposure to inorganic arsenic (As), a carcinogen and epigenetic toxicant, has been associated with lower circulating levels of insulin-like growth factor 1 (IGF1) and impaired growth in children of pre-school age. The aim of this study was to assess the potential impact of exposure to As on IGF1 and insulin-like growth factor-binding protein 3 (IGFBP3) as well as DNA methylation changes in 9-year-old children. To this end, we studied 9-year-old children from a longitudinal mother-child cohort in rural Bangladesh (n = 551). Prenatal and concurrent exposure to As was assessed via concentrations in maternal urine at gestational week 8 and in child urine at 9 years, measured by HPLC-HG-ICPMS. Plasma IGF1 and IGFBP3 concentrations were quantified with immunoassays. DNA methylation was measured in blood mononuclear cells at 9 years in a sub-sample (n = 113) using the Infinium HumanMethylation450K BeadChip. In multivariable-adjusted linear regression models, prenatal As (natural log-transformed), but not children's concurrent urinary As, was positively associated with IGFBP3 concentrations (β = 76, 95% CI 19, 133). As concentrations were not associated with IGF1. DNA methylation analysis revealed CpGs associated with both prenatal As and IGFBP3. Mediation analysis suggested that methylation of 12 CpG sites for all children was mediator of effect for the association between prenatal As and IGFBP3. We also found differentially methylated regions, generally hypermethylated, that were associated with both prenatal As and IGFBP3. In all, our study revealed that prenatal exposure to As was positively associated with IGFBP3 concentrations in children at 9 years, independent of IGF1, and this association may, at least in part, be epigenetically mediated.

DNA methylation of imprinted genes in Mexican-American newborn children with prenatal phthalate exposure

AIM

Imprinted genes exhibit expression in a parent-of-origin-dependent manner and are critical for child development. Recent limited evidence suggests that prenatal exposure to phthalates, ubiquitous endocrine disruptors, can affect their epigenetic dysregulation.

MATERIALS & METHODS

We quantified DNA methylation of nine imprinted gene differentially methylated regions by pyrosequencing in 296 cord blood DNA samples in a Mexican-American cohort. Fetal exposure was estimated by phthalate metabolite concentrations in maternal urine samples during pregnancy.

RESULTS

Several differentially methylated regions of imprinted genes were associated with high molecular weight phthalates. The most consistent, positive, and false discovery rate significant associations were observed for MEG3.

CONCLUSION

Phthalate exposure in utero may affect methylation status of imprinted genes in newborn children.

DNA-Methylation and Body Composition in Preschool Children: Epigenome-Wide-Analysis in the European Childhood Obesity Project (CHOP)-Study

Adiposity and obesity result from the interaction of genetic variation and environmental factors from very early in life, possibly mediated by epigenetic processes. Few Epigenome-Wide-Association-Studies have identified DNA-methylation (DNAm) signatures associated with BMI and body composition in children. Body composition by Bio-Impedance-Analysis and genome-wide DNAm in whole blood were assessed in 374 pre-school children from four European countries. Associations were tested by linear regression adjusted for sex, age, centre, education, 6 WBC-proportions according to Houseman and 30 principal components derived from control probes. Specific DNAm variants were identified to be associated with BMI (212), fat-mass (230), fat-free-mass (120), fat-mass-index (24) and fat-free-mass-index (15). Probes in genes SNED1(IRE-BP1), KLHL6, WDR51A(POC1A), CYTH4-ELFN2, CFLAR, PRDM14, SOS1, ZNF643(ZFP69B), ST6GAL1, C3orf70, CILP2, MLLT4 and ncRNA LOC101929268 remained significantly associated after Bonferroni-correction of P-values. We provide novel evidence linking DNAm with (i) altered lipid and glucose metabolism, (ii) diabetes and (iii) body size and composition in children. Both common and specific epigenetic signatures among measures were also revealed. The causal direction with phenotypic measures and stability of DNAm variants throughout the life course remains unclear and longitudinal analysis in other populations is required. These findings give support for potential epigenetic programming of body composition and obesity.

Newborn genome-wide DNA methylation in association with pregnancy anxiety reveals a potential role for GABBR1

BACKGROUND

There is increasing evidence for the role of prenatal stress in shaping offspring DNA methylation and disease susceptibility. In the current study, we aimed to identify genes and pathways associated with pregnancy anxiety using a genome-wide DNA methylation approach.

METHODS

We selected 22 versus 23 newborns from our Prenatal Early Life Stress (PELS) cohort, exposed to the lowest or highest degree of maternal pregnancy anxiety, respectively. Cord blood genome-wide DNA methylation was assayed using the HumanMethylation450 BeadChip (HM450, n = 45) and candidate gene methylation using EpiTYPER (n = 80). Cortisol levels were measured at 2, 4, and 12 months of age to test infant stress system (re)activity.

RESULTS

Data showed ten differentially methylated regions (DMR) when comparing newborns exposed to low versus high pregnancy anxiety scores. We validated a top DMR in the GABA-B receptor subunit 1 gene (GABBR1) revealing the association with pregnancy anxiety particularly in male newborns (most significant CpG Pearson R = 0.517, p = 0.002; average methylation Pearson R = 0.332, p = 0.039). Cord blood GABBR1 methylation was associated with infant cortisol levels in response to a routine vaccination at 4 months old.

CONCLUSIONS

In conclusion, our results show that pregnancy anxiety is associated with differential DNA methylation patterns in newborns and that our candidate gene GABBR1 is associated with infant hypothalamic-pituitary-adrenal axis response to a stressor. Our findings reveal a potential role for GABBR1 methylation in association with stress and provide grounds for further research.

Accounting for cell lineage and sex effects in the identification of cell-specific DNA methylation using a Bayesian model selection algorithm

Cell- and sex-specific differences in DNA methylation are major sources of epigenetic variation in whole blood. Heterogeneity attributable to cell type has motivated the identification of cell-specific methylation at the CpG level, however statistical methods for this purpose have been limited to pairwise comparisons between cell types or between the cell type of interest and whole blood. We developed a Bayesian model selection algorithm for the identification of cell-specific methylation profiles that incorporates knowledge of shared cell lineage and allows for the identification of differential methylation profiles in one or more cell types simultaneously. Under the proposed methodology, sex-specific differences in methylation by cell type are also assessed. Using publicly available, cell-sorted methylation data, we show that 51.3% of female CpG markers and 61.4% of male CpG markers identified were associated with differential methylation in more than one cell type. The impact of cell lineage on differential methylation was also highlighted. An evaluation of sex-specific differences revealed differences in CD56⁺NK methylation, within both single and multi- cell dependent methylation patterns. Our findings demonstrate the need to account for cell lineage in studies of differential methylation and associated sex effects.

Genetic and epigenetic regulation of YKL-40 in childhood

BACKGROUND

Circulating levels of the chitinase-like protein YKL-40 are influenced by genetic variation in its encoding gene (chitinase 3-like 1 [CHI3L1]) and are increased in patients with several diseases, including asthma. Epigenetic regulation of circulating YKL-40 early in life is unknown.

OBJECTIVE

We sought to determine (1) whether methylation levels at CHI3L1 CpG sites mediate the association of CHI3L1 single nucleotide polymorphisms (SNPs) with YKL-40 levels in the blood and (2) whether these biomarkers (CHI3L1 SNPs, methylation profiles, and YKL-40 levels) are associated with asthma in early childhood.

METHODS

We used data from up to 2405 participants from the Spanish Infancia y Medio Ambiente; the Swedish Barn/Children, Allergy, Milieu, Stockholm, Epidemiological survey; and the Dutch Prevention and Incidence of Asthma and Mite Allergy birth cohorts. Associations between 68 CHI3L1 SNPs, methylation levels at 14 CHI3L1 CpG sites in whole-blood DNA, and circulating YKL-40 levels at 4 years of age were tested by using correlation analysis, multivariable regression, and mediation analysis. Each of these biomarkers was also tested for association with asthma at 4 years of age by using multivariable logistic regression.

RESULTS

YKL-40 levels were significantly associated with 7 SNPs and with methylation at 5 CpG sites. Consistent associations between these 7 SNPs (particularly rs10399931 and rs4950928) and 5 CpG sites were observed. Alleles linked to lower YKL-40 levels were associated with higher methylation levels. Participants with high YKL-40 levels (defined as the highest YKL-40 tertile) had increased odds for asthma compared with subjects with low YKL-40 levels (meta-analyzed adjusted odds ratio, 1.90 [95% CI, 1.08-3.36]). In contrast, neither SNPs nor methylation levels at CpG sites in CHI3L1 were associated with asthma.

CONCLUSIONS

The effects of CHI3L1 genetic variation on circulating YKL-40 levels are partly mediated by methylation profiles. In our study YKL-40 levels, but not CHI3L1 SNPs or methylation levels, were associated with childhood asthma.

Prenatal phthalate exposure and altered patterns of DNA methylation in cord blood

Epigenetic changes such as DNA methylation may be a molecular mechanism through which environmental exposures affect health. Phthalates are known endocrine disruptors with ubiquitous exposures in the general population including pregnant women, and they have been linked with a number of adverse health outcomes. We examined the association between in utero phthalate exposure and altered patterns of cord blood DNA methylation in 336 Mexican-American newborns. Concentrations of 11 phthalate metabolites were analyzed in maternal urine samples collected at 13 and 26 weeks gestation as a measure of fetal exposure. DNA methylation was assessed using the Infinium HumanMethylation 450K BeadChip adjusting for cord blood cell composition. To identify differentially methylated regions (DMRs) that may be more informative than individual CpG sites, we used two different approaches, DMRcate and comb-p. Regional assessment by both methods identified 27 distinct DMRs, the majority of which were in relation to multiple phthalate metabolites. Most of the significant DMRs (67%) were observed for later pregnancy (26 weeks gestation). Further, 51% of the significant DMRs were associated with the di-(2-ethylhexyl) phthalate metabolites. Five individual CpG sites were associated with phthalate metabolite concentrations after multiple comparisons adjustment (FDR), all showing hypermethylation. Genes with DMRs were involved in inflammatory response (IRAK4 and ESM1), cancer (BRCA1 and LASP1), endocrine function (CNPY1), and male fertility (IFT140, TESC, and PRDM8). These results on differential DNA methylation in newborns with prenatal phthalate exposure provide new insights and targets to explore mechanism of adverse effects of phthalates on human health. Environ. Mol. Mutagen. 58:398-410, 2017. © 2017 Wiley Periodicals, Inc.

Cell type specific DNA methylation in cord blood: A 450K-reference data set and cell count-based validation of estimated cell type composition

Epigenome-wide association studies of prenatal exposure to different environmental factors are becoming increasingly common. These studies are usually performed in umbilical cord blood. Since blood comprises multiple cell types with specific DNA methylation patterns, confounding caused by cellular heterogeneity is a major concern. This can be adjusted for using reference data consisting of DNA methylation signatures in cell types isolated from blood. However, the most commonly used reference data set is based on blood samples from adult males and is not representative of the cell type composition in neonatal cord blood. The aim of this study was to generate a reference data set from cord blood to enable correct adjustment of the cell type composition in samples collected at birth. The purity of the isolated cell types was very high for all samples (>97.1%), and clustering analyses showed distinct grouping of the cell types according to hematopoietic lineage. We explored whether this cord blood and the adult peripheral blood reference data sets impact the estimation of cell type composition in cord blood samples from an independent birth cohort (MoBa, n = 1092). This revealed significant differences for all cell types. Importantly, comparison of the cell type estimates against matched cell counts both in the cord blood reference samples (n = 11) and in another independent birth cohort (Generation R, n = 195), demonstrated moderate to high correlation of the data. This is the first cord blood reference data set with a comprehensive examination of the downstream application of the data through validation of estimated cell types against matched cell counts.

Small-Magnitude Effect Sizes in Epigenetic End Points are Important in Children's Environmental Health Studies: The Children's Environmental Health and Disease Prevention Research Center's Epigenetics Working Group

BACKGROUND

Characterization of the epigenome is a primary interest for children's environmental health researchers studying the environmental influences on human populations, particularly those studying the role of pregnancy and early-life exposures on later-in-life health outcomes.

OBJECTIVES

Our objective was to consider the state of the science in environmental epigenetics research and to focus on DNA methylation and the collective observations of many studies being conducted within the Children's Environmental Health and Disease Prevention Research Centers, as they relate to the Developmental Origins of Health and Disease (DOHaD) hypothesis.

METHODS

We address the current laboratory and statistical tools available for epigenetic analyses, discuss methods for validation and interpretation of findings, particularly when magnitudes of effect are small, question the functional relevance of findings, and discuss the future for environmental epigenetics research.

DISCUSSION

A common finding in environmental epigenetic studies is the small-magnitude epigenetic effect sizes that result from such exposures. Although it is reasonable and necessary that we question the relevance of such small effects, we present examples in which small effects persist and have been replicated across populations and across time. We encourage a critical discourse on the interpretation of such small changes and further research on their functional relevance for children's health.

CONCLUSION

The dynamic nature of the epigenome will require an emphasis on future longitudinal studies in which the epigenome is profiled over time, over changing environmental exposures, and over generations to better understand the multiple ways in which the epigenome may respond to environmental stimuli.

Epigenetics and allergy: from basic mechanisms to clinical applications

Allergic diseases are on the rise in the Western world and well-known allergy-protecting and -driving factors such as microbial and dietary exposure, pollution and smoking mediate their influence through alterations of the epigenetic landscape. Here, we review key facts on the involvement of epigenetic modifications in allergic diseases and summarize and critically evaluate the lessons learned from epigenome-wide association studies. We show the potential of epigenetic changes for various clinical applications: as diagnostic tools, to assess tolerance following immunotherapy or possibly predict the success of therapy at an early time point. Furthermore, new technological advances such as epigenome editing and DNAzymes will allow targeted alterations of the epigenome in the future and provide novel therapeutic tools.

CpG Methylation across the adipogenic PPARγ gene and its relationship with birthweight and child BMI at 9 years

Background

To examine methylation of the peroxisome proliferator-activated receptor γ (PPARγ) gene and its relationship with child weight status, at birth and 9 years.

Methods

We measured PPARγ methylation across 23 CpG sites using the Infinium Illumina 450 k array for children from the Center for the Health Assessment of Mothers and Children of Salinas (CHAMACOS) cohort at birth (N = 373) and 9 years (N = 245).

Results

Methylation level correlation patterns across the 23 PPARγ CpG sites were conserved between birth and 9-year ages. We found high inter-CpG correlations between sites 1–3 (methylation block 1) and also between sites 18–23 (methylation block 2) for both time points, although these patterns were less pronounced at 9 years. Additionally, sites 1–3 (north shore) had the highest intra-CpG correlations over time (r = 0.24, 0.42, and 0.3; P = 0.002, P < 0.001, P < 0.001, respectively). PPARγ methylation levels tended to increase with age, and the largest differences were observed for north shore sites (7.4%). Adjusting for sex, both site 1 and site 20 (gene body) methylation at birth was significantly and inversely associated with birth weight (β = −0.13, P = 0.033; β = −0.09, P = 0.025, respectively). Similarly, we found that site 1 and site 20 methylation at 9 years was significantly and inversely associated with 9-year BMI z-score (β = −0.41, P = 0.015; β = −0.23, P = 0.045, respectively).

Conclusion

Our results indicate that PPARγ methylation is highly organized and conserved over time, and highlight the potential functional importance of north shore sites, adding to a better understanding of regional human methylome patterns. Overall, our results suggest that PPARγ methylation may be associated with child body size.

Electronic supplementary material

The online version of this article (doi:10.1186/s12881-016-0365-4) contains supplementary material, which is available to authorized users.

Genome-wide methylation data mirror ancestry information

Background

Genetic data are known to harbor information about human demographics, and genotyping data are commonly used for capturing ancestry information by leveraging genome-wide differences between populations. In contrast, it is not clear to what extent population structure is captured by whole-genome DNA methylation data.

Results

We demonstrate, using three large-cohort 450K methylation array data sets, that ancestry information signal is mirrored in genome-wide DNA methylation data and that it can be further isolated more effectively by leveraging the correlation structure of CpGs with cis-located SNPs. Based on these insights, we propose a method, EPISTRUCTURE, for the inference of ancestry from methylation data, without the need for genotype data.

Conclusions

EPISTRUCTURE can be used to infer ancestry information of individuals based on their methylation data in the absence of corresponding genetic data. Although genetic data are often collected in epigenetic studies of large cohorts, these are typically not made publicly available, making the application of EPISTRUCTURE especially useful for anyone working on public data. Implementation of EPISTRUCTURE is available in GLINT, our recently released toolset for DNA methylation analysis at: http://glint-epigenetics.readthedocs.io.

Electronic supplementary material

The online version of this article (doi:10.1186/s13072-016-0108-y) contains supplementary material, which is available to authorized users.

Training a model for estimating leukocyte composition using whole-blood DNA methylation and cell counts as reference

AIM

Whole-blood DNA methylation depends on the underlying leukocyte composition and confounding hereby is a major concern in epigenome-wide association studies. Cell counts are often missing or may not be feasible. Computational approaches estimate leukocyte composition from DNA methylation based on reference datasets of purified leukocytes. We explored the possibility to train such a model on whole-blood DNA methylation and cell counts without the need for purification.

MATERIALS & METHODS

Using whole-blood DNA methylation and corresponding five-part cell counts from 2445 participants from the London Life Sciences Prospective Population Study, a model was trained on a subset of 175 subjects and evaluated on the remaining.

RESULTS

Correlations between cell counts and estimated cell proportions were high (neutrophils 0.85, eosinophils 0.88, basophils 0.02, lymphocytes 0.84, monocytes 0.55) and estimated proportions explained more variance in whole-blood DNA methylation levels than counts.

CONCLUSION

Our model provided precise estimates for the common cell types.

miRNAs differentially expressed by next-generation sequencing in cord blood buffy coat samples of boys and girls

AIM

Differences in children's development and susceptibility to diseases and exposures have been observed by sex, yet human studies of sex differences in miRNAs are limited.

MATERIALS & METHODS

The genome-wide miRNA expression was characterized by sequencing-based EdgeSeq assay in cord blood buffy coats from 89 newborns, and 564 miRNAs were further analyzed.

RESULTS

Differential expression of most miRNAs was higher in boys. Neurodevelopment, RNA metabolism and metabolic ontology terms were enriched among miRNA targets. The majority of upregulated miRNAs (86%) validated by nCounter maintained positive-fold change values; however, only 21% reached statistical significance by false discovery rate.

CONCLUSION

Accounting for host factors like sex may improve the sensitivity of epigenetic analyses for epidemiological studies in early childhood.

Birth weight-for-gestational age is associated with DNA methylation at birth and in childhood

BACKGROUND

Both higher and lower fetal growth are associated with cardio-metabolic health later in life, suggesting that prenatal developmental programming determines long-term cardiovascular disease risk. Epigenetic mechanisms, which orchestrate fetal growth and development, may offer insight on the early programming of health and disease. We investigated whether birth weight-for-gestational is associated with DNA methylation at birth and mid-childhood, measured via the Infinium 450K array.

METHODS/RESULTS

Participants were from Project Viva, a pre-birth cohort of pregnant women and their children in Eastern Massachusetts. After exclusion of participants with maternal type 1 or 2 diabetes and gestational age <34 weeks, we used DNA methylation assays from 476 venous umbilical cord blood samples and a subset of 235 who additionally had peripheral blood samples available in mid-childhood (age 7-10 years). Among 392,918 CpG sites analyzed, birth weight-for-gestational age z-score was associated with cord blood DNA methylation at 34 CpGs (false discovery rate P < 0.05), after adjusting for maternal age, race/ethnicity, education, smoking, parity, delivery mode, pre-pregnancy BMI, gestational diabetes status, child sex, and estimated cord blood cell proportions based on a cord blood reference panel. Two of these CpGs were previously reported in epigenome-wide analyses of birth weight, and several other CpGs map to genes relevant to fetal growth and development. Namely, higher birth weight-for-gestational age was associated with higher methylation at four CpGs at the PBX1 locus (e.g., β (95% CI) for lead signal at cg06750897 = 1.9 (1.2, 2.6)), which encodes a transcription factor that regulates embryonic development. Birth weight-for-gestational age was also associated with mid-childhood blood DNA methylation at four of the 34 CpGs identified in cord blood analyses, including sites at the PBX1 locus described.

CONCLUSIONS

We identified CpG sites where birth weight-for-gestational age was associated with DNA methylation at birth, and for a subset of these sites, birth weight-for-gestational age was also associated with DNA methylation at mid-childhood.

Validation of a DNA methylation reference panel for the estimation of nucleated cells types in cord blood

Cord blood is widely used as surrogate tissue in epigenome-wide association studies of prenatal conditions. Cell type composition variation across samples can be an important confounder of epigenome-wide association studies in blood that constitute a mixture of cells. We evaluated a newly developed cord blood reference panel to impute cell type composition from DNA methylation levels, including nucleated red blood cells (nRBCs). We estimated cell type composition from 154 unique cord blood samples with available DNA methylation data as well as direct measurements of nucleated cell types. We observed high correlations between the estimated and measured composition for nRBCs (r = 0.92, R² = 0.85), lymphocytes (r = 0.77, R² = 0.58), and granulocytes (r = 0.72, R² = 0.52), and a moderate correlation for monocytes (r = 0.51, R² = 0.25) as well as relatively low root mean square errors from the residuals ranging from 1.4 to 5.4%. These results validate the use of the cord blood reference panel and highlight its utility and limitations for epidemiological studies.

Epigenetic studies in Developmental Origins of Health and Disease: pitfalls and key considerations for study design and interpretation

The field of Developmental Origins of Health and Disease (DOHaD) seeks to understand the relationships between early-life environmental exposures and long-term health and disease. Until recently, the molecular mechanisms underlying these phenomena were poorly understood; however, epigenetics has been proposed to bridge the gap between the environment and phenotype. Epigenetics involves the study of heritable changes in gene expression, which occur without changes to the underlying DNA sequence. Different types of epigenetic modifications include DNA methylation, post-translational histone modifications and non-coding RNAs. Increasingly, changes to the epigenome have been associated with early-life exposures in both humans and animal models, offering both an explanation for how the environment may programme long-term health, as well as molecular changes that could be developed as biomarkers of exposure and/or future disease. As such, epigenetic studies in DOHaD hold much promise; however, there are a number of factors which should be considered when designing and interpreting such studies. These include the impact of the genome on the epigenome, the tissue-specificity of epigenetic marks, the stability (or lack thereof) of epigenetic changes over time and the importance of associating epigenetic changes with changes in transcription or translation to demonstrate functional consequences. In this review, we discuss each of these key concepts and provide practical strategies to mitigate some common pitfalls with the aim of providing a useful guide for future epigenetic studies in DOHaD.

How to make DNA methylome wide association studies more powerful

Genome-wide association studies had a troublesome adolescence, while researchers increased statistical power, in part by increasing subject numbers. Interrogating the interaction of genetic and environmental influences raised new challenges of statistical power, which were not easily bested by the addition of subjects. Screening the DNA methylome offers an attractive alternative as methylation can be thought of as a proxy for the combined influences of genetics and environment. There are statistical challenges unique to DNA methylome data and also multiple features, which can be exploited to increase power. We anticipate the development of DNA methylome association study designs and new analytical methods, together with integration of data from other molecular species and other studies, which will boost statistical power and tackle causality. In this way, the molecular trajectories that underlie disease development will be uncovered.

Maternal phthalate exposure during pregnancy is associated with DNA methylation of LINE-1 and Alu repetitive elements in Mexican-American children

Phthalates are frequently used in personal care products and plasticizers and phthalate exposure is ubiquitous in the US population. Exposure to phthalates during critical periods in utero has been associated with a variety of adverse health outcomes but the biological mechanisms linking these exposures with disease are not well characterized. In this study, we examined the relationship of in utero phthalate exposure with repetitive element DNA methylation, an epigenetic marker of genome instability, in children from the longitudinal birth cohort CHAMACOS. Methylation of Alu and long interspersed nucleotide elements (LINE-1) was determined using pyrosequencing of bisulfite-treated DNA isolated from whole blood samples collected from newborns and 9 year old children (n=355). Concentrations of eleven phthalate metabolites were measured in urine collected from pregnant mothers at 13 and 26 weeks gestation. We found a consistent inverse association between prenatal concentrations of monoethyl phthalate, the most frequently detected urinary metabolite, with cord blood methylation of Alu repeats (β(95%CI): -0.14 (-0.28,0.00) and -0.16 (-0.31, -0.02)) for early and late pregnancy, respectively, and a similar but weaker association with LINE-1 methylation. Additionally, increases in urinary concentrations of di-(2-ethylhexyl) phthalate metabolites during late pregnancy were associated with lower levels of methylation of Alu repeats in 9 year old blood (significant p-values ranged from 0.003 to 0.03). Our findings suggest that prenatal exposure to some phthalates may influence differences in repetitive element methylation, highlighting epigenetics as a plausible biological mechanism through which phthalates may affect health.

Assisted reproductive technology alters deoxyribonucleic acid methylation profiles in bloodspots of newborn infants

OBJECTIVE

To evaluate the effect of infertility and intracytoplasmic sperm injection (ICSI) on DNA methylation of offspring.

DESIGN

Microarray analysis of DNA methylation in archived neonatal bloodspots of in vitro fertilization (IVF)/ICSI-conceived children compared with controls born to fertile and infertile parents.

SETTING

Academic research laboratory.

PATIENT(S)

Neonatal blood spots of 137 newborns conceived spontaneously, through intrauterine insemination (IUI), or through ICSI using fresh or cryopreserved (frozen) embryo transfer.

INTERVENTION(S)

None.

MAIN OUTCOME MEASURE(S)

The Illumina Infinium HumanMethylation450k BeadChip assay determined genome-wide DNA methylation. Methylation differences between conception groups were detected using a Bioconductor package, ChAMP, in conjunction with Adjacent Site Clustering (A-clustering).

RESULT(S)

The methylation profiles of assisted reproductive technology and IUI newborns were dramatically different from those of naturally (in vivo) conceived newborns. Interestingly, the profiles of ICSI-frozen (FET) and IUI infants were strikingly similar, suggesting that cryopreservation may temper some of the epigenetic aberrations induced by IVF or ICSI. The DNA methylation changes associated with IVF/ICSI culture conditions and/or parental infertility were detected at metastable epialleles, suggesting a lasting impact on a child's epigenome.

CONCLUSION(S)

Both infertility and ICSI alter DNA methylation at specific genomic loci, an effect that is mitigated to some extent by FET. The impact of assisted reproductive technology and/or fertility status on metastable epialleles in humans was uncovered. This study provides an expanded set of loci for future investigations on IVF populations.

DNA methylation of LINE-1 and Alu repetitive elements in relation to sex hormones and pubertal timing in Mexican-American children

BACKGROUND

The molecular mechanisms linking environmental exposures to earlier pubertal development are not well characterized. Epigenetics may play an important role, but data on the relationship between epigenetic marks and puberty, particularly in humans, is limited.

METHODS

We used pyrosequencing to measure Alu and long interspersed nucleotide elements (LINE-1) methylation in DNA isolated from whole blood samples collected from newborns and 9-y-old children (n = 266). Tanner staging was completed six times between ages 9 and 12 y to determine pubertal status, and hormone levels were measured in 12-y-old boys.

RESULTS

Among girls, we observed a suggestive trend of increased odds of breast and pubic hair development with higher Alu and LINE-1 methylation in 9-y-old blood, respectively. The strongest association identified was an inverse association of LINE-1 methylation in 9-y-old girls with odds of experiencing menarche by age 12 (OR (95% CI): 0.63 (0.46, 0.87); P = 0.005). We observed a consistent inverse relationship for Alu and LINE-1 methylation at 9 y with luteinizing hormone (LH), testosterone and follicle-stimulating hormone levels in boys but it was only significant between LINE-1 and LH.

CONCLUSION

DNA methylation of Alu and LINE-1 may be involved in puberty initiation and development. This relationship should be confirmed in future studies.

DNA methylation of cord blood cell types: Applications for mixed cell birth studies

Epigenome-wide association studies of disease widely use DNA methylation measured in blood as a surrogate tissue. Cell proportions can vary between people and confound associations of exposure or outcome. An adequate reference panel for estimating cell proportions from adult whole blood for DNA methylation studies is available, but an analogous cord blood cell reference panel is not yet available. Cord blood has unique cell types and the epigenetic signatures of standard cell types may not be consistent throughout the life course. Using magnetic bead sorting, we isolated cord blood cell types (nucleated red blood cells, granulocytes, monocytes, natural killer cells, B cells, CD4(+)T cells, and CD8(+)T cells) from 17 live births at Johns Hopkins Hospital. We confirmed enrichment of the cell types using fluorescence assisted cell sorting and ran DNA from the separated cell types on the Illumina Infinium HumanMethylation450 BeadChip array. After filtering, the final analysis was on 104 samples at 429,794 probes. We compared cell type specific signatures in cord to each other and methylation at 49.2% of CpG sites on the array differed by cell type (F-test P < 10(-8)). Differences between nucleated red blood cells and the remainder of the cell types were most pronounced (36.9% of CpG sites at P < 10(-8)) and 99.5% of these sites were hypomethylated relative to the other cell types. We also compared the mean-centered sorted cord profiles to the available adult reference panel and observed high correlation between the overlapping cell types for granulocytes and monocytes (both r=0.74), and poor correlation for CD8(+)T cells and NK cells (both r=0.08). We further provide an algorithm for estimating cell proportions in cord blood using the newly developed cord reference panel, which estimates biologically plausible cell proportions in whole cord blood samples.

Sex differences in DNA methylation assessed by 450 K BeadChip in newborns

BACKGROUND

DNA methylation is an important epigenetic mark that can potentially link early life exposures to adverse health outcomes later in life. Host factors like sex and age strongly influence biological variation of DNA methylation, but characterization of these relationships is still limited, particularly in young children.

METHODS

In a sample of 111 Mexican-American subjects (58 girls , 53 boys), we interrogated DNA methylation differences by sex at birth using the 450 K BeadChip in umbilical cord blood specimens, adjusting for cell composition.

RESULTS

We observed that ~3% of CpG sites were differentially methylated between girls and boys at birth (FDR P < 0.05). Of those CpGs, 3031 were located on autosomes, and 82.8% of those were hypermethylated in girls compared to boys. Beyond individual CpGs, we found 3604 sex-associated differentially methylated regions (DMRs) where the majority (75.8%) had higher methylation in girls. Using pathway analysis, we found that sex-associated autosomal CpGs were significantly enriched for gene ontology terms related to nervous system development and behavior. Among hits in our study, 35.9% had been previously reported as sex-associated CpG sites in other published human studies. Further, for replicated hits, the direction of the association with methylation was highly concordant (98.5-100%) with previous studies.

CONCLUSIONS

To our knowledge, this is the first reported epigenome-wide analysis by sex at birth that examined DMRs and adjusted for confounding by cell composition. We confirmed previously reported trends that methylation profiles are sex-specific even in autosomal genes, and also identified novel sex-associated CpGs in our methylome-wide analysis immediately after birth, a critical yet relatively unstudied developmental window.

PON1 as a model for integration of genetic, epigenetic, and expression data on candidate susceptibility genes

Recent genome- and epigenome-wide studies demonstrate that the DNA methylation is controlled in part by genetics, highlighting the importance of integrating genetic and epigenetic data. To better understand molecular mechanisms affecting gene expression, we used the candidate susceptibility gene paraoxonase 1 (PON1) as a model to assess associations of PON1 genetic polymorphisms with DNA methylation and arylesterase activity, a marker of PON1 expression. PON1 has been associated with susceptibility to obesity, cardiovascular disease, and pesticide exposure. In this study, we assessed DNA methylation in 18 CpG sites located along PON1 shores, shelves, and its CpG island in blood specimens collected from newborns and 9-year-old children participating (n = 449) in the CHAMACOS birth cohort study. The promoter polymorphism, PON1_-108 , was strongly associated with methylation, particularly for CpG sites located near the CpG island (P << 0.0005). Among newborns, these relationships were even more pronounced after adjusting for blood cell composition. We also observed significant decreases in arylesterase activity with increased methylation at the same nine CpG sites at both ages. Using causal mediation analysis, we found statistically significant indirect effects of methylation (β(95% confidence interval): 6.9(1.5, 12.4)) providing evidence that DNA methylation mediates the relationship between PON1_-108 genotype and PON1 expression. Our findings show that integration of genetic, epigenetic, and expression data can shed light on the functional mechanisms involving genetic and epigenetic regulation of candidate susceptibility genes like PON1.