Predicting DNA methylation level across human tissues

Epigenetic Clocks: Beyond Biological Age, Using the Past to Predict the Present and Future

Predicting health trajectories and accurately measuring aging processes across the human lifespan remain profound scientific challenges. Assessing the effectiveness and impact of interventions targeting aging is even more elusive, largely due to the intricate, multidimensional nature of aging-a process that defies simple quantification. Traditional biomarkers offer only partial perspectives, capturing limited aspects of the aging landscape. Yet, over the past decade, groundbreaking advancements have emerged. Epigenetic clocks, derived from DNA methylation patterns, have established themselves as powerful aging biomarkers, capable of estimating biological age and assessing aging rates across diverse tissues with remarkable precision. These clocks provide predictive insights into mortality and age-related disease risks, effectively distinguishing biological age from chronological age and illuminating enduring questions in gerontology. Despite significant progress in epigenetic clock development, substantial challenges remain, underscoring the need for continued investigation to fully unlock their potential in the science of aging.

From Mother to Child: Epigenetic Signatures of Hyperglycemia and Obesity during Pregnancy

BACKGROUND

In utero exposure to maternal hyperglycemia and obesity can trigger detrimental effects in the newborn through epigenetic programming. We aimed to assess the DNA methylation levels in the promoters of MC4R and LPL genes from maternal blood, placenta, and buccal swab samples collected in children born to mothers with and without obesity and Gestational Diabetes Mellitus (GDM).

METHODS

A total of 101 Caucasian mother-infant pairs were included in this study. Sociodemographic characteristics, clinical parameters, physical activity, and adherence to the Mediterranean diet were evaluated in the third trimester of pregnancy. Clinical parameters of the newborns were recorded at birth.

RESULTS

A negative relationship between MC4R DNA methylation on the fetal side of the GDM placenta and birth weight (r = -0.630, p = 0.011) of newborns was found. MC4R DNA methylation level was lower in newborns of GDM women (CpG1: 2.8% ± 3.0%, CpG2: 3.8% ± 3.3%) as compared to those of mothers without GDM (CpG1: 6.9% ± 6.2%, CpG2: 6.8% ± 5.6%; p < 0.001 and p = 0.0033, respectively), and it was negatively correlated with weight (r = -0.229; p = 0.035), head circumference (r = -0.236; p = 0.030), and length (r = -0.240; p = 0.027) at birth. LPL DNA methylation was higher on the fetal side of the placenta in obese patients as compared to normal-weight patients (66.0% ± 14.4% vs. 55.7% ± 15.2%, p = 0.037), and it was associated with maternal total cholesterol (r = 0.770, p = 0.015) and LDL-c (r = 0.783, p = 0.012).

CONCLUSIONS

These results support the role of maternal MC4R and LPL methylation in fetal programming and in the future metabolic health of children.

Dexamethasone induces developmental axon damage in the offspring hippocampus by activating miR-210-3p/miR-362-5p to target the aberrant expression of Sonic Hedgehog

Given the extensive application of dexamethasone in both clinical settings and the livestock industry, human exposure to this drug can occur through various sources and pathways. Prior research has indicated that prenatal exposure to dexamethasone (PDE) heightens the risk of cognitive and emotional disorders in offspring. Axonal development impairment is a frequent pathological underpinning for neuronal dysfunction in these disorders, yet it remains unclear if it plays a role in the neural damage induced by PDE in the offspring. Through RNA-seq and bioinformatics analysis, we found that various signaling pathways related to nervous system development, including axonal development, were altered in the hippocampus of PDE offspring. Among them, the Sonic Hedgehog (SHH) signaling pathway was the most significantly altered and crucial for axonal development. By using miRNA-seq and targeting miRNAs and glucocorticoid receptor (GR) expression, we identified miR-210-3p and miR-362-5p, which can target and suppress SHH expression. Their abnormal high expression was associated with GR activation in PDE fetal rats. Further testing of PDE offspring rats and infant peripheral blood samples exposed to dexamethasone in utero showed that SHH expression was significantly decreased in peripheral blood mononuclear cells (PBMCs) and was positively correlated with SHH expression in the hippocampus and the expression of the axonal development marker growth-associated protein-43. In summary, PDE-induced hippocampal GR-miR-210-3p/miR-362-5p-SHH signaling axis changes lead to axonal developmental damage. SHH expression in PBMCs may reflect axonal developmental damage in PDE offspring and could serve as a warning marker for fetal axonal developmental damage.

DNA methylation is associated with hair trace elements in female adolescents from two vulnerable populations in the Colombian Caribbean

Exposure to trace elements (TEs) influences DNA methylation patterns, which may be associated with disease development. Vulnerable populations, such as adolescents undergoing maturity, are susceptible to the effects of TE exposure. The aim of this study was to analyze the association of hair TE concentration with DNA methylation in a sample from female adolescents living in two communities in the Colombian Caribbean coast. Hair and blood samples were obtained from 45 females, between 13 and 16 years of age. Seventeen TEs were quantified in hair samples. DNA methylation was measured in leukocytes using the Infinium MethylationEPIC BeadChip. Linear models were employed to identify differentially methylated positions (DMPs) adjusting for age, body mass index, mother's education, and cell type composition. Among the tested elements, vanadium, chromium, nickel, copper, zinc, yttrium, tin, and barium were significantly associated with DMPs (false discovery rate < 0.05), registering 225, 1, 2, 184, 1, 209 189, and 104 hits, respectively. Most of the DMPs were positively associated with TEs and located in open sea regions. The greatest number of DMPs was annotated to the HOXA3 and FOXO3 genes, related to regulation of gene expression and oxidative stress, respectively. These findings suggest that DNA methylation may be involved in linking exposure to TEs among female adolescents to downstream health risks.

Blood DNA Methylation Analysis Reveals a Distinctive Epigenetic Signature of Vasospasm in Aneurysmal Subarachnoid Hemorrhage

Vasospasm is a potentially preventable cause of poor prognosis in patients with aneurysmal subarachnoid hemorrhage (aSAH). Epigenetics might provide insight on its molecular mechanisms. We aimed to analyze the association between differential DNA methylation (DNAm) and development of vasospasm. We conducted an epigenome-wide association study in 282 patients with aSAH admitted to our hospital. DNAm was assessed with the EPIC Illumina chip (> 850 K CpG sites) in whole-blood samples collected at hospital admission. We identified differentially methylated positions (DMPs) at the CpG level using Cox regression models adjusted for potential confounders, and then we used the DMP results to find differentially methylated regions (DMRs) and enriched biological pathways. A total of 145 patients (51%) experienced vasospasm. In the DMP analysis, we identified 31 CpGs associated with vasospasm at p-value < 10-5. One of them (cg26189827) was significant at the genome-wide level (p-value < 10-8), being hypermethylated in patients with vasospasm and annotated to SUGCT gene, mainly expressed in arteries. Region analysis revealed 13 DMRs, some of them annotated to interesting genes such as POU5F1, HLA-DPA1, RUFY1, and CYP1A1. Functional enrichment analysis showed the involvement of biological processes related to immunity, inflammatory response, oxidative stress, endothelial nitric oxide, and apoptosis. Our findings show, for the first time, a distinctive epigenetic signature of vasospasm in aSAH, establishing novel links with essential biological pathways, including inflammation, immune responses, and oxidative stress. Although further validation is required, our results provide a foundation for future research into the complex pathophysiology of vasospasm.

Predicting locus-specific DNA methylation levels in cancer and paracancer tissues

Aim: To predict base-resolution DNA methylation in cancerous and paracancerous tissues. Material & methods: We collected six cancer DNA methylation datasets from The Cancer Genome Atlas and five cancer datasets from Gene Expression Omnibus and established machine learning models using paired cancerous and paracancerous tissues. Tenfold cross-validation and independent validation were performed to demonstrate the effectiveness of the proposed method. Results: The developed cross-tissue prediction models can substantially increase the accuracy at more than 68% of CpG sites and contribute to enhancing the statistical power of differential methylation analyses. An XGBoost model leveraging multiple correlating CpGs may elevate the prediction accuracy. Conclusion: This study provides a powerful tool for DNA methylation analysis and has the potential to gain new insights into cancer research from epigenetics.

Tissue-specific RNA methylation prediction from gene expression data using sparse regression models

N6-methyladenosine (m6A) is a highly prevalent and conserved post-transcriptional modification observed in mRNA and long non-coding RNA (lncRNA). Identifying potential m6A sites within RNA sequences is crucial for unraveling the potential influence of the epitranscriptome on biological processes. In this study, we introduce Exp2RM, a novel approach that formulates single-site-based tissue-specific elastic net models for predicting tissue-specific methylation levels utilizing gene expression data. The resulting ensemble model demonstrates robust predictive performance for tissue-specific methylation levels, with an average R-squared value of 0.496 and a median R-squared value of 0.482 across all 22 human tissues. Since methylation distribution varies among tissues, we trained the model to incorporate similar patterns, significantly improves accuracy with the median R-squared value increasing to 0.728. Additonally, functional analysis reveals Exp2RM's ability to capture coefficient genes in relevant biological processes. This study emphasizes the importance of tissue-specific methylation distribution in enhancing prediction accuracy and provides insights into the functional implications of methylation sites.

MLACNN: an attention mechanism-based CNN architecture for predicting genome-wide DNA methylation

Methylation is an important epigenetic regulation of methylation genes that plays a crucial role in regulating biological processes. While traditional methods for detecting methylation in biological experiments are constantly improving, the development of artificial intelligence has led to the emergence of deep learning and machine learning methods as a new trend. However, traditional machine learning-based methods rely heavily on manual feature extraction, and most deep learning methods for studying methylation extract fewer features due to their simple network structures. To address this, we propose a bottomneck network based on an attention mechanism and use new methods to ensure that the deep network can learn more effective features while minimizing overfitting. This approach enables the model to learn more features from nucleotide sequences and make better predictions of methylation. The model uses three coding methods to encode the original DNA sequence and then applies feature fusion based on attention mechanisms to obtain the best fusion method. Our results demonstrate that MLACNN outperforms previous methods and achieves more satisfactory performance.

DNA Methylation and Telomeres-Their Impact on the Occurrence of Atrial Fibrillation during Cardiac Aging

Atrial fibrillation (AF) is the most common arrhythmia in humans. AF is characterized by irregular and increased atrial muscle activation. This high-frequency activation obliterates the synchronous work of the atria and ventricles, reducing myocardial performance, which can lead to severe heart failure or stroke. The risk of developing atrial fibrillation depends largely on the patient's history. Cardiovascular diseases are considered aging-related pathologies; therefore, deciphering the role of telomeres and DNA methylation (mDNA), two hallmarks of aging, is likely to contribute to a better understanding and prophylaxis of AF. In honor of Prof. Elizabeth Blackburn's 75th birthday, we dedicate this review to the discovery of telomeres and her contribution to research on aging.

Low H3K27 acetylation of SF1 in PBMC: a biomarker for prenatal dexamethasone exposure-caused adrenal insufficiency of steroid synthesis in male offspring

Dexamethasone is widely used to treat pregnancy disorders related to premature delivery. However, lots of researches have confirmed that prenatal dexamethasone exposure (PDE) could increase the risk of offspring multiple diseases. This study was designed to elucidate the epigenetic mechanism of adrenal developmental programming and explore its early warning marker in peripheral blood mononuclear cells (PBMC). We found the adrenal morphological and functional changes of PDE male offspring rats before and after birth, which were mainly performed as the decreased serum corticosterone concentration, steroidogenic acute regulatory (StAR) protein expression, and histone 3 lysine 27 acetylation (H3K27ac) level of steroidogenic factor 1 (SF1) promoter region and its expression. Simultaneously, the expressions of glucocorticoid receptor (GR) and histone acetylation enzyme 5 (HDAC5) in the PDE male fetal rats were increased. In vitro, dexamethasone reduced the expression of SF1, StAR, and cortisol production and still increased the expression of GR and HDAC5, the binding between GR and SF1 promoter region, and protein interaction between GR and HDAC5. GR siRNA or HDAC5 siRNA was able to reverse the above roles of dexamethasone. Furthermore, in vivo, we confirmed that H3K27ac levels of SF1 promoter region and its expression in PBMC of the PDE group were decreased before and after birth, showing a positive correlation with the same indexes in adrenal. Meanwhile, in clinical trials, we confirmed that prenatal dexamethasone application decreased H3K27ac of SF1 promoter region and its expression in neonatal PBMC. In conclusion, PDE-caused adrenal insufficiency of male offspring rats was related to adrenal GR activated by dexamethasone in uterus. The activated GR, on the one hand, increased its direct binding to SF1 promoter region to inhibit its expression, on the other hand, upregulated and recruited HDAC5 to decrease H3K27ac level of SF1 promoter region, and strengthened the inhibition of SF1 and subsequent StAR expression.

DNA methylation-based biomarkers for cardiometabolic-related traits and their importance for risk stratification

Recent findings

The prevalence of cardiometabolic syndrome in adults is increasing worldwide, highlighting the importance of biomarkers for individuals' classification based on their health status. Although cardiometabolic risk scores and diagnostic criteria have been developed aggregating adverse health effects of individual conditions on the overall syndrome, none of them has gained unanimous acceptance. Therefore, novel molecular biomarkers have been developed to better understand the risk, onset and progression of both individual conditions and the overall cardiometabolic syndrome.

Summary

Consistent associations between whole blood DNA methylation (DNAm) levels at several single genomic (i.e. CpG) sites and both individual and aggregated cardiometabolic conditions supported the creation of second-generation DNAm-based cardiometabolic-related biomarkers. These biomarkers linearly combine individual DNAm levels from key CpG sites, selected by a two-step machine learning procedures. They can be used, even retrospectively, in populations with extant whole blood DNAm levels and without observed cardiometabolic phenotypes.

Purpose of review

Here we offer an overview of the second-generation DNAm-based cardiometabolic biomarkers, discussing methodological advancements and implications on the interpretation and generalizability of the findings. We finally emphasize the contribution of DNAm-based biomarkers for risk stratification beyond traditional factors and discuss limitations and future directions of the field.

Completing Single-Cell DNA Methylome Profiles via Transfer Learning Together With KL-Divergence

The high level of sparsity in methylome profiles obtained using whole-genome bisulfite sequencing in the case of low biological material amount limits its value in the study of systems in which large samples are difficult to assemble, such as mammalian preimplantation embryonic development. The recently developed computational methods for addressing the sparsity by imputing missing have their limits when the required minimum data coverage or profiles of the same tissue in other modalities are not available. In this study, we explored the use of transfer learning together with Kullback-Leibler (KL) divergence to train predictive models for completing methylome profiles with very low coverage (below 2%). Transfer learning was used to leverage less sparse profiles that are typically available for different tissues for the same species, while KL divergence was employed to maximize the usage of information carried in the input data. A deep neural network was adopted to extract both DNA sequence and local methylation patterns for imputation. Our study of training models for completing methylome profiles of bovine oocytes and early embryos demonstrates the effectiveness of transfer learning and KL divergence, with individual increase of 29.98 and 29.43%, respectively, in prediction performance and 38.70% increase when the two were used together. The drastically increased data coverage (43.80-73.6%) after imputation powers downstream analyses involving methylomes that cannot be effectively done using the very low coverage profiles (0.06-1.47%) before imputation.

Targeting Epigenetics and Non-coding RNAs in Myocardial Infarction: From Mechanisms to Therapeutics

Myocardial infarction (MI) is a complicated pathology triggered by numerous environmental and genetic factors. Understanding the effect of epigenetic regulation mechanisms on the cardiovascular disease would advance the field and promote prophylactic methods targeting epigenetic mechanisms. Genetic screening guides individualised MI therapies and surveillance. The present review reported the latest development on the epigenetic regulation of MI in terms of DNA methylation, histone modifications, and microRNA-dependent MI mechanisms and the novel therapies based on epigenetics.

Integrating SNPs-based genetic risk factor with blood epigenomic response of differentially arsenic-exposed rural subjects reveals disease-associated signaling pathways

Arsenic (As) contamination in groundwater is responsible for numerous adverse health outcomes among millions of people. Epigenetic alterations are among the most widely studied mechanisms of As toxicity. To understand how As exposure alters gene expression through epigenetic modifications, a systematic genome-wide study was designed to address the impact of multiple important single nucleotide polymorphisms (SNPs) related to As exposure on the methylome of drinking water As-exposed rural subjects from Pakistan. Urinary As levels were used to stratify subjects into low, medium and high exposure groups. Genome-wide DNA methylation was investigated using MeDIP in combination with NimbleGen 2.1 M Deluxe Promotor arrays. Transcriptome levels were measured using Agilent 8 × 60 K expression arrays. Genotyping of selected SNPs (As3MT, DNMT1a, ERCC2, EGFR and MTHFR) was measured and an integrated genetic risk factor for each respondent was calculated by assigning a specific value to the measured genotypes based on known risk allele numbers. To select a representative model related to As exposure we compared 9 linear mixed models comprising of model 1 (including the genetic risk factor), model 2 (without the genetic risk factor) and models with individual SNPs incorporated into the methylome data. Pathway analysis was performed using ConsensusPathDB. Model 1 comprising the integrated genetic risk factor disclosed biochemical pathways including muscle contraction, cardio-vascular diseases, ATR signaling, GPCR signaling, methionine metabolism and chromatin modification in association with hypo- and hyper-methylated gene targets. A unique pathway (direct P53 effector) was found associated with the individual DNMT1a polymorphism due to hyper-methylation of CSE1L and TRRAP. Most importantly, we provide here the first evidence of As-associated DNA methylation in relation with gene expression of ATR, ATF7IP, TPM3, UBE2J2. We report the first evidence that integrating SNPs data with methylome data generates a more representative epigenome profile and discloses a better insight in disease risks of As-exposed individuals.

DNA Methylation Imputation Across Platforms

In this chapter, we will provide a review on imputation in the context of DNA methylation, specifically focusing on a penalized functional regression (PFR) method we have previously developed. We will start with a brief review of DNA methylation, genomic and epigenomic contexts where imputation has proven beneficial in practice, and statistical or computational methods proposed for DNA methylation in the recent literature (Subheading 1). The rest of the chapter (Subheadings 2-4) will provide a detailed review of our PFR method proposed for across-platform imputation, which incorporates nonlocal information using a penalized functional regression framework. Subheading 2 introduces commonly employed technologies for DNA methylation measurement and describes the real dataset we have used in the development of our method: the acute myeloid leukemia (AML) dataset from The Cancer Genome Atlas (TCGA) project. Subheading 3 comprehensively reviews our method, encompassing data harmonization prior to model building, the actual building of penalized functional regression model, post-imputation quality filter, and imputation quality assessment. Subheading 4 shows the performance of our method in both simulation and the TCGA AML dataset, demonstrating that our penalized functional regression model is a valuable across-platform imputation tool for DNA methylation data, particularly because of its ability to boost statistical power for subsequent epigenome-wide association study. Finally, Subheading 5 provides future perspectives on imputation for DNA methylation data.

DNA methylation profile of genes involved in inflammation and autoimmunity correlates with vascular function in morbidly obese adults

Obesity is a major risk factor for cardiovascular disease. Blood-detected epigenetic profiles may serve as non-invasive clinically relevant biomarkers. Therefore, we investigated DNA methylation of genes involved in inflammation in peripheral blood of obese subjects and lean controls and their correlation with cardiometabolic measurements. We obtained blood and adipose tissue (AT) samples from bariatric patients (n = 24) and control adults (n = 24). AT-isolated arterioles were tested for flow-induced dilation (FID) and production of nitric oxide (NO) and reactive oxygen species (ROS). Brachial artery flow-mediated dilation (FMD) was measured via doppler ultrasound. Promoter methylation of 94 genes involved in inflammation and autoimmunity were analysed in whole-blood DNA in relation to vascular function and cardiometabolic risk factors. 77 genes had ahigher methylated fraction in the controls compare obese subjects and 28 proinflammatory genes were significantly hypomethylated in the obese individuals; on top of these genes are CXCL1, CXCL12, CXCL6, IGF2BP2, HDAC4, IL12A, and IL17RA. Fifteen of these genes had significantly higher mRNA in obese subjects compared to controls; on top of these genes are CXCL6, TLR5, IL6ST, EGR1, IL15RA, and HDAC4. Methylation % inversely correlated with BMI, total fat %, visceral fat%, blood pressure, fasting plasma insulin, serum IL6 and C-reactive protein, arteriolar ROS, and alcohol consumption and positive correlations with lean %, HDL, plasma folate and vitamin B12, arteriolar FID and NO production, and brachial FMD. Our results suggest that vascular dysfunction in obese adults may be attributed to asystemic hypomethylation and over expression of the immune-related genes.

Predicting environmentally responsive transgenerational differential DNA methylated regions (epimutations) in the genome using a hybrid deep-machine learning approach

BACKGROUND

Deep learning is an active bioinformatics artificial intelligence field that is useful in solving many biological problems, including predicting altered epigenetics such as DNA methylation regions. Deep learning (DL) can learn an informative representation that addresses the need for defining relevant features. However, deep learning models are computationally expensive, and they require large training datasets to achieve good classification performance.

RESULTS

One approach to addressing these challenges is to use a less complex deep learning network for feature selection and Machine Learning (ML) for classification. In the current study, we introduce a hybrid DL-ML approach that uses a deep neural network for extracting molecular features and a non-DL classifier to predict environmentally responsive transgenerational differential DNA methylated regions (DMRs), termed epimutations, based on the extracted DL-based features. Various environmental toxicant induced epigenetic transgenerational inheritance sperm epimutations were used to train the model on the rat genome DNA sequence and use the model to predict transgenerational DMRs (epimutations) across the entire genome.

CONCLUSION

The approach was also used to predict potential DMRs in the human genome. Experimental results show that the hybrid DL-ML approach outperforms deep learning and traditional machine learning methods.

A 6-CpG validated methylation risk score model for metabolic syndrome: The HyperGEN and GOLDN studies

There has been great interest in genetic risk prediction using risk scores in recent years, however, the utility of scores developed in European populations and later applied to non-European populations has not been successful. The goal of this study was to create a methylation risk score (MRS) for metabolic syndrome (MetS), demonstrating the utility of MRS across race groups using cross-sectional data from the Hypertension Genetic Epidemiology Network (HyperGEN, N = 614 African Americans (AA)) and the Genetics of Lipid Lowering Drugs and Diet Network (GOLDN, N = 995 European Americans (EA)). To demonstrate this, we first selected cytosine-guanine dinucleotides (CpG) sites measured on Illumina Methyl450 arrays previously reported to be significantly associated with MetS and/or component conditions in more than one race/ethnic group (CPT1A cg00574958, PHOSPHO1 cg02650017, ABCG1 cg06500161, SREBF1 cg11024682, SOCS3 cg18181703, TXNIP cg19693031). Second, we calculated the parameter estimates for the 6 CpGs in the HyperGEN data (AA) and used the beta estimates as weights to construct a MRS in HyperGEN (AA), which was validated in GOLDN (EA). We performed association analyses using logistic mixed models to test the association between the MRS and MetS, adjusting for covariates. Results showed the MRS was significantly associated with MetS in both populations. In summary, a MRS for MetS was a strong predictor for the condition across two race groups, suggesting MRS may be useful to examine metabolic disease risk or related complications across race/ethnic groups.

Integrative, multi-omics, analysis of blood samples improves model predictions: applications to cancer

Background

Cancer genomic studies often include data collected from several omics platforms. Each omics data source contributes to the understanding of the underlying biological process via source specific (“individual”) patterns of variability. At the same time, statistical associations and potential interactions among the different data sources can reveal signals from common biological processes that might not be identified by single source analyses. These common patterns of variability are referred to as “shared” or “joint”. In this work, we show how the use of joint and individual components can lead to better predictive models, and to a deeper understanding of the biological process at hand. We identify joint and individual contributions of DNA methylation, miRNA and mRNA expression collected from blood samples in a lung cancer case–control study nested within the Norwegian Women and Cancer (NOWAC) cohort study, and we use such components to build prediction models for case–control and metastatic status. To assess the quality of predictions, we compare models based on simultaneous, integrative analysis of multi-source omics data to a standard non-integrative analysis of each single omics dataset, and to penalized regression models. Additionally, we apply the proposed approach to a breast cancer dataset from The Cancer Genome Atlas.

Results

Our results show how an integrative analysis that preserves both components of variation is more appropriate than standard multi-omics analyses that are not based on such a distinction. Both joint and individual components are shown to contribute to a better quality of model predictions, and facilitate the interpretation of the underlying biological processes in lung cancer development.

Conclusions

In the presence of multiple omics data sources, we recommend the use of data integration techniques that preserve the joint and individual components across the omics sources. We show how the inclusion of such components increases the quality of model predictions of clinical outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12859-021-04296-0.

CUE: CpG impUtation ensemble for DNA methylation levels across the human methylation450 (HM450) and EPIC (HM850) BeadChip platforms

DNA methylation at CpG dinucleotides is one of the most extensively studied epigenetic marks. With technological advancements, geneticists can profile DNA methylation with multiple reliable approaches. However, profiling platforms can differ substantially in the CpGs they assess, consequently hindering integrated analysis across platforms. Here, we present CpG impUtation Ensemble (CUE), which leverages multiple classical statistical and modern machine learning methods, to impute from the Illumina HumanMethylation450 (HM450) BeadChip to the Illumina HumanMethylationEPIC (HM850) BeadChip. Data were analysed from two population cohorts with methylation measured both by HM450 and HM850: the Extremely Low Gestational Age Newborns (ELGAN) study (n = 127, placenta) and the VA Boston Posttraumatic Stress Disorder (PTSD) genetics repository (n = 144, whole blood). Cross-validation results show that CUE achieves the lowest predicted root-mean-square error (RMSE) (0.026 in PTSD) and the highest accuracy (99.97% in PTSD) compared with five individual methods tested, including k-nearest-neighbours, logistic regression, penalized functional regression, random forest, and XGBoost. Finally, among all 339,033 HM850-only CpG sites shared between ELGAN and PTSD, CUE successfully imputed 289,604 (85.4%) sites, where success was defined as RMSE < 0.05 and accuracy >95% in PTSD. In summary, CUE is a valuable tool for imputing CpG methylation from the HM450 to HM850 platform.

Methylome Patterns of Cattle Adaptation to Heat Stress

Heat stress has a detrimental impact on cattle health, welfare and productivity by affecting gene expression, metabolism and immune response, but little is known on the epigenetic mechanisms mediating the effect of temperature at the cellular and organism level. In this study, we investigated genome-wide DNA methylation in blood samples collected from 5 bulls of the heat stress resilient Nellore breed and 5 bulls of the Angus that are more heat stress susceptible, exposed to the sun and high temperature-high humidity during the summer season of the Brazilian South-East region. The methylomes were analyzed during and after the exposure by Reduced Representation Bisulfite Sequencing, which provided genome-wide single-base resolution methylation profiles. Significant methylation changes between stressful and recovery periods were observed in 819 genes. Among these, 351 were only seen in Angus, 366 were specific to Nellore, and 102 showed significant changes in methylation patterns in both breeds. KEGG and Gene Ontology (GO) enrichment analyses showed that responses were breed-specific. Interestingly, in Nellore significant genes and pathways were mainly involved in stress responses and cellular defense and were under methylated during heat stress, whereas in Angus the response was less focused. These preliminary results suggest that heat challenge induces changes in methylation patterns in specific loci, which should be further scrutinized to assess their role in heat tolerance.

Ultra performance liquid chromatography-tandem mass spectrometry assay for the quantification of RNA and DNA methylation

Previous studies have reported that nucleic acid methylation is a critical element in cardiovascular disease, and most studies mainly focused on sequencing and biochemical research. Here we developed an Ultra performance liquid chromatography-tandem mass spectrometry (UPLC-MS/ MS) method for the quantification analysis of the dissociative epigenetic modified nucleosides (5mdC, 5mrC, m⁶A) in Myocardial Infarction (MI) SD rats from different periods (1 week, 4 weeks, 8 weeks) after the surgery. The samples for analysis were obtained from heart tissue and blood of the rats. All the quantification results are compared with the sham-operated group. Total RNA and DNA were isolated by enzymatic hydrolytic methods before the UPLC-MS/MS analysis. The statistical analysis demonstrates the dynamic changes of modified nucleosides in MI rats, and it showed good specificity, accuracy, stability and less samples were needed in the method. In this paper, we discovered that the concentration of 5mdC, 5mrC, m⁶A from heart tissue significantly increased at 8 weeks after the surgery. Furthermore, UPLC-MS/MS helps us observe the similar change of the concentration of those 3 methylated biomarkers in peripheral blood after 8 weeks. The result shows that the dynamic process of those 3 methylated biomarkers in peripheral blood is related to the content of methylated biomarkers from the heart tissue. Based on the scientific evidence available, we proved that the methylation of genetic materials in peripheral blood is similar to myocardial infarction tissue. The relation between them indicates that peripheral blood could be a promising alternative to the heart tissue which monitor the level of methylation and MI diagnosis-aided.

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

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

A novel computational strategy for DNA methylation imputation using mixture regression model (MRM)

BACKGROUND

DNA methylation is an important heritable epigenetic mark that plays a crucial role in transcriptional regulation and the pathogenesis of various human disorders. The commonly used DNA methylation measurement approaches, e.g., Illumina Infinium HumanMethylation-27 and -450 BeadChip arrays (27 K and 450 K arrays) and reduced representation bisulfite sequencing (RRBS), only cover a small proportion of the total CpG sites in the human genome, which considerably limited the scope of the DNA methylation analysis in those studies.

RESULTS

We proposed a new computational strategy to impute the methylation value at the unmeasured CpG sites using the mixture of regression model (MRM) of radial basis functions, integrating information of neighboring CpGs and the similarities in local methylation patterns across subjects and across multiple genomic regions. Our method achieved a better imputation accuracy over a set of competing methods on both simulated and empirical data, particularly when the missing rate is high. By applying MRM to an RRBS dataset from subjects with low versus high bone mineral density (BMD), we recovered methylation values of ~ 300 K CpGs in the promoter regions of chromosome 17 and identified some novel differentially methylated CpGs that are significantly associated with BMD.

CONCLUSIONS

Our method is well applicable to the numerous methylation studies. By expanding the coverage of the methylation dataset to unmeasured sites, it can significantly enhance the discovery of novel differential methylation signals and thus reveal the mechanisms underlying various human disorders/traits.

Genetics and Epigenetics of Atrial Fibrillation

Atrial fibrillation (AF) is known to be the most common supraventricular arrhythmia affecting up to 1% of the general population. Its prevalence exponentially increases with age and could reach up to 8% in the elderly population. The management of AF is a complex issue that is addressed by extensive ongoing basic and clinical research. AF centers around different types of disturbances, including ion channel dysfunction, Ca²⁺-handling abnormalities, and structural remodeling. Genome-wide association studies (GWAS) have uncovered over 100 genetic loci associated with AF. Most of these loci point to ion channels, distinct cardiac-enriched transcription factors, as well as to other regulatory genes. Recently, the discovery of post-transcriptional regulatory mechanisms, involving non-coding RNAs (especially microRNAs), DNA methylation, and histone modification, has allowed to decipher how a normal heart develops and which modifications are involved in reshaping the processes leading to arrhythmias. This review aims to provide a current state of the field regarding the identification and functional characterization of AF-related epigenetic regulatory networks

Genome-Wide DNA Methylation and Gene Expression Profiles in Cows Subjected to Different Stress Level as Assessed by Cortisol in Milk

Dairy cattle health, wellbeing and productivity are deeply affected by stress. Its influence on metabolism and immune response is well known, but the underlying epigenetic mechanisms require further investigation. In this study, we compared DNA methylation and gene expression signatures between two dairy cattle populations falling in the high- and low-variant tails of the distribution of milk cortisol concentration (MC), a neuroendocrine marker of stress in dairy cows. Reduced Representation Bisulfite Sequencing was used to obtain a methylation map from blood samples of these animals. The high and low groups exhibited similar amounts of methylated CpGs, while we found differences among non-CpG sites. Significant methylation changes were detected in 248 genes. We also identified significant fold differences in the expression of 324 genes. KEGG and Gene Ontology (GO) analysis showed that genes of both groups act together in several pathways, such as nervous system activity, immune regulatory functions and glucocorticoid metabolism. These preliminary results suggest that, in livestock, cortisol secretion could act as a trigger for epigenetic regulation and that peripheral changes in methylation can provide an insight into central nervous system functions.

Genome-wide investigation of DNA methylation in congenital adrenal hyperplasia

Patients with congenital adrenal hyperplasia (CAH) are at risk of long-term cognitive and metabolic sequelae with some of the effects being attributed to the chronic glucocorticoid treatment that they receive. Our pilot study investigates genome-wide DNA methylation in patients with CAH to determine whether there is preliminary evidence for epigenomic reprogramming as well as any relationship to patient outcome. Here, we analysed CD4 + T cell DNA from 28 patients with CAH (mean age = 18.5 ± 6.5 years [y]) and 37 population controls (mean age = 17.0 ± 6.1 y) with the Infinium-HumanMethylation450 BeadChip array to measure genome-wide locus-specific DNA methylation levels. Effects of CAH, phenotype and CYP21A2 genotype on methylation were investigated as well as the association between differentially methylated CpGs and glucose homeostasis, blood lipid profile, and cognitive functions. In addition, we report data on a small cohort of 11 patients (mean age = 19.1, ±6.0 y) with CAH who were treated prenatally with dexamethasone (DEX) in addition to postnatal glucocorticoid treatment. We identified two CpGs to be associated with patient phenotype: cg18486102 (located in the FAIM2 gene; rho = 0.58, adjusted p = 0.027) and cg02404636 (located in the SFI1 gene; rho = 0.58, adjusted p = 0.038). cg02404636 was also associated with genotype (rho = 0.59, adjusted p = 0.024). Higher levels of serum C-peptide was also observed in patients with CAH (p = 0.044). Additionally, levels of C-peptide and HbA1c were positively correlated with patient phenotype (p = 0.044 and p = 0.034) and genotype (p = 0.044 and p = 0.033), respectively. No significant association was found between FAIM2 methylation and cognitive or metabolic outcome. However, SFI1 TSS methylation was associated with fasting plasma HDL cholesterol levels (p = 0.035). In conclusion, in this pilot study, higher methylation levels in CpG sites covering FAIM2 and SFI1 were associated with disease severity. Hypermethylation in these genes may have implications for long-term cognitive and metabolic outcome in patients with CAH, although the data must be interpreted with caution due to the small sample size. Additional studies in larger cohorts are therefore warranted.

Methylation Data Processing Protocol and Comparison of Blood and Cerebral Spinal Fluid Following Aneurysmal Subarachnoid Hemorrhage

One challenge in conducting DNA methylation-based epigenome-wide association study (EWAS) is the appropriate cleaning and quality-checking of data to minimize biases and experimental artifacts, while simultaneously retaining potential biological signals. These issues are compounded in studies that include multiple tissue types, and/or tissues for which reference data are unavailable to assist in adjusting for cell-type mixture, for example cerebral spinal fluid (CSF). For our study that evaluated blood and CSF taken from aneurysmal subarachnoid hemorrhage (aSAH) patients, we developed a protocol to clean and quality-check genome-wide methylation levels and compared the methylomic profiles of the two tissues to determine whether blood is a suitable surrogate for CSF. CSF samples were collected from 279 aSAH patients longitudinally during the first 14 days of hospitalization, and a subset of 88 of these patients also provided blood samples within the first 2 days. Quality control (QC) procedures included identification and exclusion of poor performing samples and low-quality probes, functional normalization, and correction for cell-type heterogeneity via surrogate variable analysis (SVA). Significant differences in rates of poor sample performance was observed between blood (1.1% failing QC) and CSF (9.12% failing QC; p = 0.003). Functional normalization increased the concordance of methylation values among technical replicates in both CSF and blood. SVA improved the asymptotic behavior of the test of association in a simulated EWAS under the null hypothesis. To determine the suitability of blood as a surrogate for CSF, we calculated the correlations of adjusted methylation values at each CpG between blood and CSF globally and by genomic regions. Overall, mean within-CpG correlation was low (r < 0.26), suggesting that blood is not a suitable surrogate for global methylation in CSF. However, differences in the magnitude of the correlation were observed by genomic region (CpG island, shore, shelf, open sea; p < 0.001 for all) and orientation with respect to nearby genes (3' UTR, transcription start site, exon, body, 5' UTR; p < 0.01 for all). In conclusion, the correlation analysis and QC pipelines indicated that DNA extracted from blood was not, overall, a suitable surrogate for DNA from CSF in aSAH methylomic studies.

A genome-wide analysis of DNA methylation identifies a novel association signal for Lp(a) concentrations in the LPA promoter

Lipoprotein(a) [Lp(a)] is a major cardiovascular risk factor, which is largely genetically determined by one major gene locus, the LPA gene. Many aspects of the transcriptional regulation of LPA are poorly understood and the role of epigenetics has not been addressed yet. Therefore, we conducted an epigenome-wide analysis of DNA methylation on Lp(a) levels in two population-based studies (total n = 2208). We identified a CpG site in the LPA promoter which was significantly associated with Lp(a) concentrations. Surprisingly, the identified CpG site was found to overlap the SNP rs76735376. We genotyped this SNP de-novo in three studies (total n = 7512). The minor allele of rs76735376 (1.1% minor allele frequency) was associated with increased Lp(a) values (p = 1.01e-59) and explained 3.5% of the variation of Lp(a). Statistical mediation analysis showed that the effect on Lp(a) is rather originating from the base change itself and is not mediated by DNA methylation levels. This finding is supported by eQTL data from 208 liver tissue samples from the GTEx project, which shows a significant association of the rs76735376 minor allele with increased LPA expression. To evaluate, whether the association signal at rs76735376 may actually be derived from a stronger eQTL signal in LD with this SNP, eQTL association results of all correlated SNPs (r2≥0.1) were integrated with genetic association results. This analysis pinpointed to rs10455872 as the potential trigger of the effect of rs76735376. Furthermore, both SNPs coincide with short apo(a) isoforms. Adjusting for both, rs10455872 and the apo(a) isoforms diminished the effect size of rs76735376 to 5.38 mg/dL (p = 0.0463). This indicates that the effect of rs76735376 can be explained by both an independent effect of the SNP and a strong correlation with rs10455872 and apo(a) isoforms.

DNA methylation in promoter regions of genes involved in the reproductive and metabolic function of children born to women with PCOS

Clinical and experimental evidences indicate that epigenetic modifications induced by the prenatal environment are related to metabolic and reproductive derangements in polycystic ovary syndrome (PCOS). Alterations in the leptin and adiponectin systems, androgen signalling and antimüllerian hormone (AMH) levels have been observed in PCOS women and in their offspring. Using a targeted Next-Generation Sequencing (NGS), we studied DNA methylation in promoter regions of the leptin (LEP), leptin receptor (LEPR), adiponectin (ADIPOQ), adiponectin receptor 1 and 2 (ADIPOR1 and ADIPOR2), AMH and androgen receptor (AR) genes in 24 sons and daughters of women with PCOS (12 treated with metformin during pregnancy) and 24 children born to non-PCOS women during early infancy (2-3 months of age). Genomic DNA was extracted from whole blood, bisulphite converted and sequenced by NGS. Girls showed differences between groups in 1 CpG site of LEPR, 2 of LEP, 1 of ADIPOR2 and 2 of AR. Boys showed differences in 5 CpG sites of LEP, 3 of AMH and 9 of AR. Maternal metformin treatment prevented some of these changes in LEP, ADIPOR2 and partially in AR in girls, and in LEP and AMH in boys. Maternal BMI at early pregnancy was inversely correlated with the methylation levels of the ChrX-67544981 site in the whole group of girls (r = -0.530, p = 0.008) and with the global Z-score in all boys (r = -0.539, p = 0.007). These data indicate that the intrauterine PCOS environment predisposes the offspring to acquire certain sex-dependent DNA methylation patterns in the promoter regions of metabolic and reproductive genes.

SRHiC: A Deep Learning Model to Enhance the Resolution of Hi-C Data

Hi-C data is important for studying chromatin three-dimensional structure. However, the resolution of most existing Hi-C data is generally coarse due to sequencing cost. Therefore, it will be helpful if we can predict high-resolution Hi-C data from low-coverage sequencing data. Here we developed a novel and simple computational method based on deep learning named super-resolution Hi-C (SRHiC) to enhance the resolution of Hi-C data. We verified SRHiC on Hi-C data in human cell line. We also evaluated the generalization power of SRHiC by enhancing Hi-C data resolution in other human and mouse cell types. Results showed that SRHiC outperforms the state-of-the-art methods in accuracy of prediction.

Environmental temperature and human epigenetic modifications: A systematic review

The knowledge about the effects of environmental temperature on human epigenome is a potential key to understand the health impacts of temperature and to guide acclimation under climate change. We performed a systematic review on the epidemiological studies that have evaluated the association between environmental temperature and human epigenetic modifications. We identified seven original articles on this topic published between 2009 and 2019, including six cohort studies and one cross-sectional study. They focused on DNA methylation in elderly people (blood sample) or infants (placenta sample), with sample size ranging from 306 to 1798. These studies were conducted in relatively low temperature setting (median/mean temperature: 0.8-13 °C), and linear models were used to evaluate temperature-DNA methylation association over short period (≤28 days). It has been reported that short-term ambient temperature could affect global human DNA methylation. A total of 15 candidate genes (ICAM-1, CRAT, F3, TLR-2, iNOS, ZKSCAN4, ZNF227, ZNF595, ZNF597, ZNF668, CACNA1H, AIRE, MYEOV2, NKX1-2 and CCDC15) with methylation status associated with ambient temperature have been identified. DNA methylation on ZKSCAN4, ICAM-1 partly mediated the effect of short-term cold temperature on high blood pressure and ICAM-1 protein (related to cardiovascular events), respectively. In summary, epidemiological evidence about the impacts of environment temperature on human epigenetics remains scarce and limited to short-term linear effect of cold temperature on DNA methylation in elderly people and infants. More studies are needed to broaden our understanding of temperature related epigenetic changes, especially under a changing climate.

The Joint Analysis of Multi-Omics Data Revealed the Methylation-Expression Regulations in Atrial Fibrillation

Atrial fibrillation (AF) is one of the most prevalent heart rhythm disorder. The causes of AF include age, male sex, diabetes, hypertension, valve disease, and systolic/diastolic dysfunction. But on molecular level, its mechanisms are largely unknown. In this study, we collected 10 patients with persistent atrial fibrillation, 10 patients with paroxymal atrial fibrillation and 10 healthy individuals and did Methylation EPICBead Chip and RNA sequencing. By analyzing the methylation and gene expression data using machine learning based feature selection method Boruta, we identified the key genes that were strongly associated with AF and found their interconnections. The results suggested that the methylation of KIF15 may regulate the expression of PSMC3, TINAG, and NUDT6. The identified AF associated methylation-expression regulations may help understand the molecular mechanisms of AF from a multi-omics perspective.

Methylome-wide association study provides evidence of particulate matter air pollution-associated DNA methylation

BACKGROUND

DNA methylation (DNAm) may contribute to processes that underlie associations between air pollution and poor health. Therefore, our objective was to evaluate associations between DNAm and ambient concentrations of particulate matter (PM) ≤2.5, ≤10, and 2.5-10 μm in diameter (PM2.5; PM10; PM2.5-10).

METHODS

We conducted a methylome-wide association study among twelve cohort- and race/ethnicity-stratified subpopulations from the Women's Health Initiative and the Atherosclerosis Risk in Communities study (n = 8397; mean age: 61.5 years; 83% female; 45% African American; 9% Hispanic/Latino American). We averaged geocoded address-specific estimates of daily and monthly mean PM concentrations over 2, 7, 28, and 365 days and 1 and 12 months before exams at which we measured leukocyte DNAm in whole blood. We estimated subpopulation-specific, DNAm-PM associations at approximately 485,000 Cytosine-phosphate-Guanine (CpG) sites in multi-level, linear, mixed-effects models. We combined subpopulation- and site-specific estimates in fixed-effects, inverse variance-weighted meta-analyses, then for associations that exceeded methylome-wide significance and were not heterogeneous across subpopulations (P < 1.0 × 10-7; PCochran's Q > 0.10), we characterized associations using publicly accessible genomic databases and attempted replication in the Cooperative Health Research in the Region of Augsburg (KORA) study.

RESULTS

Analyses identified significant DNAm-PM associations at three CpG sites. Twenty-eight-day mean PM10 was positively associated with DNAm at cg19004594 (chromosome 20; MATN4; P = 3.33 × 10-8). One-month mean PM10 and PM2.5-10 were positively associated with DNAm at cg24102420 (chromosome 10; ARPP21; P = 5.84 × 10-8) and inversely associated with DNAm at cg12124767 (chromosome 7; CFTR; P = 9.86 × 10-8). The PM-sensitive CpG sites mapped to neurological, pulmonary, endocrine, and cardiovascular disease-related genes, but DNAm at those sites was not associated with gene expression in blood cells and did not replicate in KORA.

CONCLUSIONS

Ambient PM concentrations were associated with DNAm at genomic regions potentially related to poor health among racially, ethnically and environmentally diverse populations of U.S. women and men. Further investigation is warranted to uncover mechanisms through which PM-induced epigenomic changes may cause disease.

Maternal Intake of n-3 Polyunsaturated Fatty Acids During Pregnancy Is Associated With Differential Methylation Profiles in Cord Blood White Cells

A healthy diet during pregnancy is pivotal for the offspring health at birth and later in life. N-3 polyunsaturated fatty acids (n-3 PUFAs) are not endogenously produced in humans and are exclusively derived from the diet. They are pivotal for the fetus growth and neuronal development and seem beneficial in reducing the risk of cardiometabolic diseases and preventing later allergic disorders in the offspring by modulating the inflammatory immune response. In the present study, we investigated the association between maternal intakes of n-3PUFAs, profiled on maternal erythrocyte membranes at pregnancy term, and offspring DNA methylation on cord blood mononuclear cells in a sample of 118 mother–newborn pairs randomly drawn from the “Feeding fetus’ low-grade inflammation and insulin-resistance” study cohort. N-3 PUFA content on erythrocyte membranes is a validated biomarker to measure objectively medium term intake of n-3 PUFAs. Based on distribution of n-3 PUFA in the whole cohort of mothers, we identified mothers with low (n-3 PUFA concentration <25th percentile), medium (n-3 PUFAs between 25th and 75th percentiles), and high n-3 PUFA content (>75th percentile). The HumanMethylation450 BeadChip (Illumina) was used for the epigenome-wide association study using the Infinium Methylation Assay. The overall DNA methylation level was not different between the three groups while there was significant difference in methylation levels at certain sites. Indeed, 8,503 sites had significantly different methylations between low and high n-3 PUFA groups, 12,716 between low and medium n-3 PUFA groups, and 18,148 between high and medium n-3 PUFA groups. We found differentially methylated genes that belong prevalently to pathways of signal transduction, metabolism, downstream signaling of G protein-coupled receptors, and gene expression. Within these pathways, we identified four differentially methylated genes, namely, MSTN, IFNA13, ATP8B3, and GABBR2, that are involved in the onset of insulin resistance and adiposity, innate immune response, phospholipid translocation across cell membranes, and mechanisms of addiction to high fat diet, alcohol, and sweet taste. In conclusion, findings of this preliminary investigation suggest that maternal intake of n-3 PUFAs during pregnancy has potential to influence the offspring DNA methylation. Validation of results in a larger cohort and investigation of biological significance and impact on the phenotype are warranted.

Using NGS-methylation profiling to understand the molecular pathogenesis of young MI patients who have subsequent cardiac events

Globally, ischaemic heart disease is a major contributor to premature morbidity and mortality. A significant number of young Myocardial Infarction (MI) patients (aged <55 y) have subsequent cardiac events within a year of their index event. This study used Next Generation Sequencing (NGS) methylation to understand the pathogenesis in this subset of young MI patients, comparing them to a cohort of patients without recurrent events. Cases and controls were matched for age, gender, ethnicity, and comorbidities. Differential methylation analyses were performed on Reduced Representation Bisulphite Sequencing (RRBS) data. Across the group and within case-control pairs' variation were analysed. Pairwise comparisons across each matched case-control pair resulted in a list of genes that were consistently significantly differentially methylated between all 16 matched pairs. This gene list was input into pathway analysis databases. Of particular relevance to cardiac pathology the following pathways were identified as over-represented in the patients with recurrent events; cell adhesion, transcription regulation and cardiac electrical conduction, specifically relating to calcium channel activity. This study looked at methylation differences between two populations of young MI patients. There were significantly different methylation profiles between the two groups studied; key pathways were identified as specifically affected in the patients with recurrent cardiac events. Matched pairwise comparisons and detailed interpretations of DNA methylation data may help to elucidate complex pathogeneses within and between clinical subtypes. Further analysis will determine whether these epigenomic differences can be useful as predictive biomarkers of clinical progression.

MRCNN: a deep learning model for regression of genome-wide DNA methylation

Background

Determination of genome-wide DNA methylation is significant for both basic research and drug development. As a key epigenetic modification, this biochemical process can modulate gene expression to influence the cell differentiation which can possibly lead to cancer. Due to the involuted biochemical mechanism of DNA methylation, obtaining a precise prediction is a considerably tough challenge. Existing approaches have yielded good predictions, but the methods either need to combine plenty of features and prerequisites or deal with only hypermethylation and hypomethylation.

Results

In this paper, we propose a deep learning method for prediction of the genome-wide DNA methylation, in which the Methylation Regression is implemented by Convolutional Neural Networks (MRCNN). Through minimizing the continuous loss function, experiments show that our model is convergent and more precise than the state-of-art method (DeepCpG) according to results of the evaluation. MRCNN also achieves the discovery of de novo motifs by analysis of features from the training process.

Conclusions

Genome-wide DNA methylation could be evaluated based on the corresponding local DNA sequences of target CpG loci. With the autonomous learning pattern of deep learning, MRCNN enables accurate predictions of genome-wide DNA methylation status without predefined features and discovers some de novo methylation-related motifs that match known motifs by extracting sequence patterns.

Electronic supplementary material

The online version of this article (10.1186/s12864-019-5488-5) contains supplementary material, which is available to authorized users.

Locus-specific DNA methylation prediction in cord blood and placenta

DNA methylation is known to be responsive to prenatal exposures, which may be a part of the mechanism linking early developmental exposures to future chronic diseases. Many studies use blood to measure DNA methylation, yet we know that DNA methylation is tissue specific. Placenta is central to fetal growth and development, but it is rarely feasible to collect this tissue in large epidemiological studies; on the other hand, cord blood samples are more accessible. In this study, based on paired samples of both placenta and cord blood tissues from 169 individuals, we investigated the methylation concordance between placenta and cord blood. We then employed a machine-learning-based model to predict locus-specific DNA methylation levels in placenta using DNA methylation levels in cord blood. We found that methylation correlation between placenta and cord blood is lower than other tissue pairs, consistent with existing observations that placenta methylation has a distinct pattern. Nonetheless, there are still a number of CpG sites showing robust association between the two tissues. We built prediction models for placenta methylation based on cord blood data and documented a subset of 1,012 CpG sites with high correlation between measured and predicted placenta methylation levels. The resulting list of CpG sites and prediction models could help to reveal the loci where internal or external influences may affect DNA methylation in both placenta and cord blood, and provide a reference data to predict the effects on placenta in future study even when the tissue is not available in an epidemiological study.

Cardiovascular Programming During and After Diabetic Pregnancy: Role of Placental Dysfunction and IUGR

Intrauterine growth restriction (IUGR) is a condition whereby a fetus is unable to achieve its genetically determined potential size. IUGR is a global health challenge due to high mortality and morbidity amongst affected neonates. It is a multifactorial condition caused by maternal, fetal, placental, and genetic confounders. Babies born of diabetic pregnancies are usually large for gestational age but under certain conditions whereby prolonged uncontrolled hyperglycemia leads to placental dysfunction, the outcome of the pregnancy is an intrauterine growth restricted fetus with clinical features of malnutrition. Placental dysfunction leads to undernutrition and hypoxia, which triggers gene modification in the developing fetus due to fetal adaptation to adverse utero environmental conditions. Thus, in utero gene modification results in future cardiovascular programming in postnatal and adult life. Ongoing research aims to broaden our understanding of the molecular mechanisms and pathological pathways involved in fetal programming due to IUGR. There is a need for the development of effective preventive and therapeutic strategies for the management of growth-restricted infants. Information on the mechanisms involved with in utero epigenetic modification leading to development of cardiovascular disease in adult life will increase our understanding and allow the identification of susceptible individuals as well as the design of targeted prevention strategies. This article aims to systematically review the latest molecular mechanisms involved in the pathogenesis of IUGR in cardiovascular programming. Animal models of IUGR that used nutrient restriction and hypoxia to mimic the clinical conditions in humans of reduced flow of nutrients and oxygen to the fetus will be discussed in terms of cardiac remodeling and epigenetic programming of cardiovascular disease. Experimental evidence of long-term fetal programming due to IUGR will also be included.

Epigenetic Alterations Associated With Early Prenatal Dexamethasone Treatment

Prenatal treatment with dexamethasone (DEX) reduces virilization in girls with congenital adrenal hyperplasia (CAH). It has potential short- and long-term risks and has been shown to affect cognitive functions. Here, we investigate whether epigenetic modification of DNA during early developmental stages may be a key mediating mechanism by which prenatal DEX treatment could result in poor outcomes in the offspring. We analyzed genome-wide CD4+ T cell DNA methylation, assessed using the Infinium HumanMethylation450 BeadChip array in 29 individuals (mean age = 16.4 ± 5.9 years) at risk for CAH and treated with DEX during the first trimester and 37 population controls (mean age = 17.0 years, SD = 6.1 years). We identified 9672 differentially methylated probes (DMPs) associated with DEX treatment and 7393 DMPs associated with a DEX × sex interaction. DMPs were enriched in intergenic regions located near epigenetic markers for active enhancers. Functional enrichment of DMPs was mostly associated with immune functioning and inflammation but also with nonimmune-related functions. DEX-associated DMPs enriched near single nucleotide polymorphisms (SNPs) associated with inflammatory bowel disease, and DEX × sex-associated DMPs enriched near SNPs associated with asthma. DMPs in genes involved in the regulation and maintenance of methylation and steroidogenesis were identified as well. Methylation in the BDNF, FKBP5, and NR3C1 genes were associated with the performance on several Wechsler Adult Intelligence Scale-Fourth Edition subscales. In conclusion, this study indicates that DNA methylation is altered after prenatal DEX treatment. This finding may have implications for the future health of the exposed individual.

Early post-conception maternal cortisol, children’s HPAA activity and DNA methylation profiles

The hypothalamic–pituitary–adrenal axis (HPAA) plays a critical role in the functioning of all other biological systems. Thus, studying how the environment may influence its ontogeny is paramount to understanding developmental origins of health and disease. The early post-conceptional (EPC) period could be particularly important for the HPAA as the effects of exposures on organisms’ first cells can be transmitted through all cell lineages. We evaluate putative relationships between EPC maternal cortisol levels, a marker of physiologic stress, and their children’s pre-pubertal HPAA activity (n=22 dyads). Maternal first-morning urinary (FMU) cortisol, collected every-other-day during the first 8 weeks post-conception, was associated with children’s FMU cortisol collected daily around the start of the school year, a non-experimental challenge, as well as salivary cortisol responses to an experimental challenge (all Ps<0.05), with some sex-related differences. We investigated whether epigenetic mechanisms statistically mediated these links and, therefore, could provide cues as to possible biological pathways involved. EPC cortisol was associated with >5% change in children’s buccal epithelial cells’ DNA methylation for 867 sites, while children’s HPAA activity was associated with five CpG sites. Yet, no CpG sites were related to both, EPC cortisol and children’s HPAA activity. Thus, these epigenetic modifications did not statistically mediate the observed physiological links. Larger, prospective peri-conceptional cohort studies including frequent bio-specimen collection from mothers and children will be required to replicate our analyses and, if our results are confirmed, identify biological mechanisms mediating the statistical links observed between maternal EPC cortisol and children’s HPAA activity.

DNA methylation patterns in peripheral blood mononuclear cells from Holstein cattle with variable milk yield

Background

Milk yield for Holstein cows has doubled over five decades due to genetic selection and changes to management, but the molecular mechanisms that facilitated this increase are mostly unknown. Epigenetic modifications to the cattle genome are a plausible molecular mechanism to cause variation in milk yield and our objective was to describe genome-wide DNA methylation patterns in peripheral blood mononuclear cells (PBMC) from mature Holstein dairy cows with variable milk yield.

Results

Whole genome MeDIP-seq was performed following DNA extraction from PBMC of 6 lactating dairy cows from 4 different herds that varied in milk yield from 13,556 kg to 23,105 kg per 305 day lactation. We describe methylation across the genome and for 13,677 protein coding genes. Repetitive element reads were primarily mapped to satellite (36.4%), SINE (29.1%), and LINE (23.7%) regions and the majority (78.4%) of CpG sites were sequenced at least once. DNA methylation was generally low upstream of genes with the nadir occurring 95 bp prior to the transcription start site (TSS). Methylation was lower in the first exon than in later exons, was highest for introns near the intron-exon junctions, and declined downstream as the distance from the gene increased. We identified 72 differentially methylated regions (DMR) between high milk yield cows and their control, and 252 DMR across herd environments.

Conclusions

This reference methylome for cattle with extreme variation in milk yield phenotype provides a resource to more fully evaluate relationships between DNA methylation and phenotype in populations subject to selection. The detection of DMR in cows of varying milk yield suggests potential to exploit epigenetic variation in cattle improvement programs.

Electronic supplementary material

The online version of this article (10.1186/s12864-018-5124-9) contains supplementary material, which is available to authorized users.

Data for GAW20: genome-wide DNA sequence variation and epigenome-wide DNA methylation before and after fenofibrate treatment in a family study of metabolic phenotypes

GAW20 provided participants with an opportunity to comprehensively examine genetic and epigenetic variation among related individuals in the context of drug treatment response. GAW20 used data from 188 families (N = 1105) participating in the Genetics of Lipid Lowering Drugs and Diet Network (GOLDN) study (clinicaltrials.gov identifier NCT00083369), which included CD4+ T-cell DNA methylation at 463,995 cytosine-phosphate-guanine (CpG) sites measured before and after a 3-week treatment with fenofibrate, single-nucleotide variation at 906,600 loci, metabolic syndrome components ascertained before and after the drug intervention, and relevant covariates. All GOLDN participants were of European descent, with an average age of 48 years. In addition, approximately half were women and approximately 40% met the diagnostic criteria for metabolic syndrome. Unique advantages of the GAW20data set included longitudinal (3 weeks apart) measurements of DNA methylation, the opportunity to explore the contributions of both genotype and DNA methylation to the interindividual variability in drug treatment response, and the familial relationships between study participants. The principal disadvantage of GAW20/GOLDN data was the spurious correlation between batch effects and fenofibrate effects on methylation, which arose because the pre- and posttreatment methylation data were generated and normalized separately, and any attempts to remove time-dependent technical artifacts would also remove biologically meaningful changes brought on by fenofibrate. Despite this limitation, the GAW20 data set offered informative, multilayered omics data collected in a large population-based study of common disease traits, which resulted in creative approaches to integration and analysis of inherited human variation.

DNA methylation of a novel PAK4 locus influences ototoxicity susceptibility following cisplatin and radiation therapy for pediatric embryonal tumors

Background

Ototoxicity is a common adverse side effect of platinum chemotherapy and cranial radiation therapy; however, individual susceptibility is highly variable. Therefore, our objective was to conduct an epigenome-wide association study to identify differentially methylated cytosine-phosphate-guanine (CpG) sites associated with ototoxicity susceptibility among cisplatin-treated pediatric patients with embryonal tumors.

Methods

Samples were collected for a discovery cohort (n = 62) and a replication cohort (n = 18) of medulloblastoma and primitive neuroectodermal tumor patients. Posttreatment audiograms were evaluated using the International Society of Paediatric Oncology (SIOP) Boston Ototoxicity Scale. Genome-wide associations between CpG methylation and ototoxicity were examined using multiple linear regression, controlling for demographic and treatment factors.

Results

The mean cumulative dose of cisplatin was 330 mg/m2 and the mean time from end of therapy to the last available audiogram was 6.9 years. In the discovery analysis of 435233 CpG sites, 6 sites were associated with ototoxicity grade (P < 5 × 10-5) after adjusting for confounders. Differential methylation at the top CpG site identified in the discovery cohort (cg14010619, PAK4 gene) was replicated (P = 0.029) and reached genome-wide significance (P = 2.73 × 10-8) in a combined analysis. These findings were robust to a sensitivity analysis evaluating other potential confounders.

Conclusions

We identified and replicated a novel CpG methylation loci (cg14010619) associated with ototoxicity severity. Methylation at cg14010619 may modify PAK4 activity, which has been implicated in cisplatin resistance in malignant cell lines.

DIRECTION: a machine learning framework for predicting and characterizing DNA methylation and hydroxymethylation in mammalian genomes

Motivation

5-Methylcytosine and 5-Hydroxymethylcytosine in DNA are major epigenetic modifications known to significantly alter mammalian gene expression. High-throughput assays to detect these modifications are expensive, labor-intensive, unfeasible in some contexts and leave a portion of the genome unqueried. Hence, we devised a novel, supervised, integrative learning framework to perform whole-genome methylation and hydroxymethylation predictions in CpG dinucleotides. Our framework can also perform imputation of missing or low quality data in existing sequencing datasets. Additionally, we developed infrastructure to perform in silico, high-throughput hypotheses testing on such predicted methylation or hydroxymethylation maps.

Results

We test our approach on H1 human embryonic stem cells and H1-derived neural progenitor cells. Our predictive model is comparable in accuracy to other state-of-the-art DNA methylation prediction algorithms. We are the first to predict hydroxymethylation in silico with high whole-genome accuracy, paving the way for large-scale reconstruction of hydroxymethylation maps in mammalian model systems. We designed a novel, beam-search driven feature selection algorithm to identify the most discriminative predictor variables, and developed a platform for performing integrative analysis and reconstruction of the epigenome. Our toolkit DIRECTION provides predictions at single nucleotide resolution and identifies relevant features based on resource availability. This offers enhanced biological interpretability of results potentially leading to a better understanding of epigenetic gene regulation.

Availability and implementation

http://www.pradiptaray.com/direction, under CC-by-SA license.

Contacts

pradiptaray@gmail.com or mchen@utdallas.edu or michael.zhang@utdallas.edu.

Supplementary information

Supplementary data are available at Bioinformatics online.

BoostMe accurately predicts DNA methylation values in whole-genome bisulfite sequencing of multiple human tissues

Background

Bisulfite sequencing is widely employed to study the role of DNA methylation in disease; however, the data suffer from biases due to coverage depth variability. Imputation of methylation values at low-coverage sites may mitigate these biases while also identifying important genomic features associated with predictive power.

Results

Here we describe BoostMe, a method for imputing low-quality DNA methylation estimates within whole-genome bisulfite sequencing (WGBS) data. BoostMe uses a gradient boosting algorithm, XGBoost, and leverages information from multiple samples for prediction. We find that BoostMe outperforms existing algorithms in speed and accuracy when applied to WGBS of human tissues. Furthermore, we show that imputation improves concordance between WGBS and the MethylationEPIC array at low WGBS depth, suggesting improved WGBS accuracy after imputation.

Conclusions

Our findings support the use of BoostMe as a preprocessing step for WGBS analysis.

Electronic supplementary material

The online version of this article (10.1186/s12864-018-4766-y) contains supplementary material, which is available to authorized users.

A closer look at cross-validation for assessing the accuracy of gene regulatory networks and models

Cross-validation (CV) is a technique to assess the generalizability of a model to unseen data. This technique relies on assumptions that may not be satisfied when studying genomics datasets. For example, random CV (RCV) assumes that a randomly selected set of samples, the test set, well represents unseen data. This assumption doesn’t hold true where samples are obtained from different experimental conditions, and the goal is to learn regulatory relationships among the genes that generalize beyond the observed conditions. In this study, we investigated how the CV procedure affects the assessment of supervised learning methods used to learn gene regulatory networks (or in other applications). We compared the performance of a regression-based method for gene expression prediction estimated using RCV with that estimated using a clustering-based CV (CCV) procedure. Our analysis illustrates that RCV can produce over-optimistic estimates of the model’s generalizability compared to CCV. Next, we defined the ‘distinctness’ of test set from training set and showed that this measure is predictive of performance of the regression method. Finally, we introduced a simulated annealing method to construct partitions with gradually increasing distinctness and showed that performance of different gene expression prediction methods can be better evaluated using this method.