Epigenetic Marks Research Articles

Transglutaminase 2-mediated histone monoaminylation and its role in cancer.

Transglutaminase 2 (TGM2) has been known as a well-characterized factor regulating the progression of multiple types of cancer, due to its multifunctional activities and the ubiquitous signaling pathways it is involved in. As a member of the transglutaminase family, TGM2 catalyzes protein post-translational modifications (PTMs), including monoaminylation, amide hydrolysis, cross-linking, etc., through the transamidation of variant glutamine-containing protein substrates. Recent discoveries revealed histone as an important category of TGM2 substrates, thus identifying histone monoaminylation as an emerging epigenetic mark, which is highly enriched in cancer cells and possesses significant regulatory functions of gene transcription. In this review, we will summarize recent advances in TGM2-mediated histone monoaminylation as well as its role in cancer and discuss the key research methodologies to better understand this unique epigenetic mark, thereby shedding light on the therapeutic potential of TGM2 as a druggable target in cancer treatment.

Deep5hmC: predicting genome-wide 5-hydroxymethylcytosine landscape via a multimodal deep learning model.

5-Hydroxymethylcytosine (5hmC), a crucial epigenetic mark with a significant role in regulating tissue-specific gene expression, is essential for understanding the dynamic functions of the human genome. Despite its importance, predicting 5hmC modification across the genome remains a challenging task, especially when considering the complex interplay between DNA sequences and various epigenetic factors such as histone modifications and chromatin accessibility. Using tissue-specific 5hmC sequencing data, we introduce Deep5hmC, a multimodal deep learning framework that integrates both the DNA sequence and epigenetic features such as histone modification and chromatin accessibility to predict genome-wide 5hmC modification. The multimodal design of Deep5hmC demonstrates remarkable improvement in predicting both qualitative and quantitative 5hmC modification compared to unimodal versions of Deep5hmC and state-of-the-art machine learning methods. This improvement is demonstrated through benchmarking on a comprehensive set of 5hmC sequencing data collected at four developmental stages during forebrain organoid development and across 17 human tissues. Compared to DeepSEA and random forest, Deep5hmC achieves close to 4% and 17% improvement of Area Under the Receiver Operating Characteristic (AUROC) across four forebrain developmental stages, and 6% and 27% across 17 human tissues for predicting binary 5hmC modification sites; and 8% and 22% improvement of Spearman correlation coefficient across four forebrain developmental stages, and 17% and 30% across 17 human tissues for predicting continuous 5hmC modification. Notably, Deep5hmC showcases its practical utility by accurately predicting gene expression and identifying differentially hydroxymethylated regions (DhMRs) in a case-control study of Alzheimer's disease (AD). Deep5hmC significantly improves our understanding of tissue-specific gene regulation and facilitates the development of new biomarkers for complex diseases. Deep5hmC is available via https://github.com/lichen-lab/Deep5hmC.

Global Characterization of DNA Methylation during Rice Leaf Angle Development.

During plant development and growth, genomic DNA accumulates chemical markers that determine the levels of gene expression. DNA methylation is an important epigenetic marker involved in plant developmental events. However, the characterization of the role of DNA methylation in rice leaf angle development has lagged behind. Herein, we performed bisulfite sequencing to characterize DNA methylation sites and performed transcriptome and small RNA sequencing during leaf angle development. The results revealed a global reduction in CG methylation during leaf angle establishment. A reduction in gene body CG methylation appears to play a vital role in leaf angle development. The hypomethylated and weakly expressed genes were functionally enriched in the brassinosteroid and auxin signaling pathways. Additionally, the main DNA methyltransferases were inactive. The addition of exogenous DNA methylation inhibitor 5-azacytidine increased the leaf angle, which confirmed that DNA methylation is crucial for leaf angle development. This study revealed a gradual decrease in 24-nucleotide siRNA levels during leaf angle development, particularly in relation to the enrichment of 24-nucleotide siRNAs at different hypomethylated regions that induce leaf angle inclination. Our results indicate crucial roles for DNA methylation in the rice leaf angle developmental stages.

Kappa Opioid Receptor Activation Induces Epigenetic Silencing of Brain-Derived Neurotropic Factor via HDAC5 in Depression.

Treatment-resistant depression (TRD) occurs in almost 50% of the depressed patients. Central kappa opioid receptor (KOR) agonism has been demonstrated to induce depression and anxiety, while KOR antagonism alleviates depression-like symptoms in rodent models and TRD in clinical studies. Previously, we have shown that sustained KOR activation leads to a TRD-like phenotype in mice, and modulation of brain-derived neurotrophic factor (BDNF) expression in the prefrontal cortex (PFC) appears to be one of the molecular determinants of the antidepressant response. In the present study, we observed that sustained KOR activation by a selective agonist, U50488, selectively reduced the levels of Bdnf transcripts II, IV, and Bdnf CDS (protein-coding Exon IX) in the PFC and cultured primary cortical neurons, which was blocked by selective KOR antagonist, norbinaltorphimine. Considering the crucial role of epigenetic pathways in BDNF expression, we further investigated the role of various epigenetic markers in KOR-induced BDNF downregulation in mice. We observed that treatment with U50488 resulted in selective and specific downregulation of acetylation at the ninth lysine residue of the histone H3 protein (H3K9ac) and upregulation of histone deacetylase 5 (HDAC5) expression in the PFC. Further, using anti-H3K9ac and anti-HDAC5 antibodies in the chromatin immune precipitation assay, we detected decreased enrichment of H3K9ac and increased HDAC5 binding at Bdnf II and IV transcripts after U50488 treatment, which were blocked by a selective KOR antagonist, norbinaltorphimine. Further mechanistic studies using HDAC5 selective inhibitor, LMK235, in primary cortical neurons and adeno-associated viral shRNA-mediated HDAC5-knockdown in the PFC of mice demonstrated an essential role of HDAC5 in KOR-mediated reduction of Bdnf expression in the PFC and in depression-like symptoms in mice. These results suggest that KOR engages multiple pathways to induce depression-like symptoms in mice and provide novel insights into the mechanisms by which activation of KOR regulates major depressive disorders.

Deep Learning Sequence Models for Transcriptional Regulation.

Deciphering the regulatory code of gene expression and interpreting the transcriptional effects of genome variation are critical challenges in human genetics. Modern experimental technologies have resulted in an abundance of data, enabling the development of sequence-based deep learning models that link patterns embedded in DNA to the biochemical and regulatory properties contributing to transcriptional regulation, including modeling epigenetic marks, 3D genome organization, and gene expression, with tissue and cell-type specificity. Such methods can predict the functional consequences of any noncoding variant in the human genome, even rare or never-before-observed variants, and systematically characterize their consequences beyond what is tractable from experiments or quantitative genetics studies alone. Recently, the development and application of interpretability approaches have led to the identification of key sequence patterns contributing to the predicted tasks, providing insights into the underlying biological mechanisms learned and revealing opportunities for improvement in future models.

Gastroenteropancreatic neuroendocrine tumors: Epigenetic landscape and clinical implications

Neuroendocrine tumors (NETs) are a rare, heterogenous group of neoplasms arising from cells of the neuroendocrine system. Amongst solid tumor malignancies, NETs are notable for overall genetic stability and recent data supports the notion that epigenetic changes may drive NET pathogenesis. In this review, major epigenetic mechanisms of NET pathogenesis are reviewed, including changes in DNA methylation, histone modification, chromatin remodeling, and microRNA. Prognostic implications of the above are discussed, as well as the expanding diagnostic utility of epigenetic markers in NETs. Lastly, preclinical and clinical evaluations of epigenetically targeted therapies in NETs and are reviewed, with a focus on future directions in therapeutic advancement.

β-Aminobutyric acid promotes stress tolerance, physiological adjustments, as well as broad epigenetic changes at DNA and RNA nucleobases in field elms (Ulmus minor)

Backgroundβ-Aminobutyric acid (BABA) has been successfully used to prime stress resistance in numerous plant species; however, its effectiveness in forest trees has been poorly explored thus far. This study aimed to investigate the influence of BABA on morphological, physiological, and epigenetic parameters in field elms under various growth conditions. Epigenetic changes were assessed in both DNA and RNA through the use of reversed-phase ultra-performance liquid chromatography (UPLC) coupled with sensitive mass spectrometry.ResultsThe presented results confirm the influence of BABA on the development, physiology, and stress tolerance in field elms. However, the most important findings are related to the broad epigenetic changes promoted by this amino acid, which involve both DNA and RNA. Our findings confirm, for the first time, that BABA influences not only well-known epigenetic markers in plants, such as 5-methylcytosine, but also several other non-canonical nucleobases, such as 5-hydroxymethyluracil, 5-formylcytosine, 5-hydroxymethylcytosine, N6-methyladenine, uracil (in DNA) and thymine (in RNA). The significant effect on the levels of N6-methyladenine, the main bacterial epigenetic marker, is particularly noteworthy. In this case, the question arises as to whether this effect is due to epigenetic changes in the microbiome, the plant genome, or both.ConclusionsThe plant phenotype is the result of complex interactions between the plant’s DNA, the microbiome, and the environment. We propose that different types of epigenetic changes in the plant and microbiome may play important roles in the largely unknown memory process that enables plants to adapt faster to changing environmental conditions.

Alterations in DNA 5-hydroxymethylation patterns in the hippocampus of an experimental model of chronic epilepsy

Temporal lobe epilepsy (TLE) is a type of focal epilepsy characterized by spontaneous recurrent seizures originating from the hippocampus. The epigenetic reprogramming hypothesis of epileptogenesis suggests that the development of TLE is associated with alterations in gene transcription changes resulting in a hyperexcitable network in TLE. DNA 5-methylcytosine (5-mC) is an epigenetic mechanism that has been associated with chronic epilepsy. However, the contribution of 5-hydroxymethylcytosine (5-hmC), a product of 5-mC demethylation by the Ten-Eleven Translocation (TET) family proteins in chronic TLE is poorly understood. 5-hmC is abundant in the brain and acts as a stable epigenetic mark altering gene expression through several mechanisms. Here, we found that the levels of bulk DNA 5-hmC but not 5-mC were significantly reduced in the hippocampus of human TLE patients and in the kainic acid (KA) TLE rat model. Using 5-hmC hMeDIP-sequencing, we characterized 5-hmC distribution across the genome and found bidirectional regulation of 5-hmC at intergenic regions within gene bodies. We found that hypohydroxymethylated 5-hmC intergenic regions were associated with several epilepsy-related genes, including Gal, SV2, and Kcnj11 and hyperdroxymethylation 5-hmC intergenic regions were associated with Gad65, TLR4, and Bdnf gene expression. Mechanistically, Tet1 knockdown in the hippocampus was sufficient to decrease 5-hmC levels and increase seizure susceptibility following KA administration. In contrast, Tet1 overexpression in the hippocampus resulted in increased 5-hmC levels associated with improved seizure resiliency in response to KA. These findings suggest an important role for 5-hmC as an epigenetic regulator of epilepsy that can be manipulated to influence seizure outcomes.

Chronic ethanol exposure decreases H3K27me3 in the Il6 promoter region of macrophages and generates persistent dysfunction on neutrophils during fungal infection.

The aim of this study was to investigate the effects of ethanol exposure on epigenetic markers in bone marrow (BM) and their impact on inflammatory response during Aspergillus fumigatus infection. Chronic ethanol exposure decreased H3K27me3 enrichment in the Il6 promoter region while increased H3K4me3 enrichment in Tnf. Chimeric mice were generated by transplanting BM from mice exposed to ethanol or water. Infection of ethanol-chimeric mice culminated in higher clinical scores, although there was no effect on mortality. However, previous chronic exposure to ethanol affects persistently the inflammatory response in lung tissue, demonstrated by increased lung damage, neutrophil accumulation and IL-6, TNF and CXCL2 production in ethanol-chimeric mice, resulting in a decreased neutrophil infiltration into the alveolar space. Neutrophil killing and phagocytosis were also significantly lower. Moreover, BM derived macrophages (BMDM) from ethanol-chimeric mice stimulated with A. fumigatus conidia exhibited higher levels of TNF, CXCL2 and IL-6 release and a higher killing activity. The Il6 promoter of BMDM from ethanol-chimeric mice exhibited a reduction in H3K27me3 enrichment, a finding also observed in BM donors exposed to ethanol. These evidences demonstrate that prior chronic alcohol exposure of bone-marrow modify immune effector cells functions impairing the inflammatory response during A. fumigatus infection. These findings highlight the persistent impact of chronic ethanol exposure on infectious disease outcomes.

DNA methylation at cannabinoid type 1 and dopamine D2 receptor genes in saliva samples of psychotic subjects: Is there an effect of Cannabis use?

Psychosis is a characterizing feature of many mental disorders that dramatically affects human thoughts and perceptions, influencing the ability to distinguish between what is real and what is not. Both genetic and environmental factors, such as stressful events or drug use, play a pivotal role in the development of symptomatology and therefore changes in the epigenome may be of relevance in modeling a psychotic phenotype. According to the well-documented dysregulation of endocannabinoid and dopaminergic system genes in schizophrenia, we investigated DNA methylation cannabinoid type 1 receptor (CNR1) and dopamine D2 receptor (DRD2) genes in saliva samples from psychotic subjects using pyrosequencing. The epigenetic mark was significantly higher and directly correlated for both genes in psychotic subjects compared to healthy controls. We also showed that these DNA methylation levels were lower in psychotic subjects reporting current delta-9-tetrahydrocannabinol (THC) consumption, a well-known risk factor for developing psychosis throughout the lifespan, resembling those of controls at least for the DRD2 gene. Overall, our data confirm the key role of CNR1 and DRD2 gene regulation in psychosis and suggest DNA methylation levels at specific CpG sites as potential biomarkers, but just in those psychotic subjects not consuming THC.

An integrated structural model of the DNA damage-responsive H3K4me3 binding WDR76:SPIN1 complex with the nucleosome

Serial capture affinity purification (SCAP) is a powerful method to isolate a specific protein complex. When combined with cross-linking mass spectrometry and computational approaches, one can build an integrated structural model of the isolated complex. Here, we applied SCAP to dissect a subpopulation of WDR76 in complex with SPIN1, a histone reader that recognizes trimethylated histone H3 lysine4 (H3K4me3). In contrast to a previous SCAP analysis of the SPIN1:SPINDOC complex, histones and the H3K4me3 mark were enriched with the WDR76:SPIN1 complex. Next, interaction network analysis of copurifying proteins and microscopy analysis revealed a potential role of the WDR76:SPIN1 complex in the DNA damage response. Since we detected 149 pairs of cross-links between WDR76, SPIN1, and histones, we then built an integrated structural model of the complex where SPIN1 recognized the H3K4me3 epigenetic mark while interacting with WDR76. Finally, we used the powerful Bayesian Integrative Modeling approach as implemented in the Integrative Modeling Platform to build a model of WDR76 and SPIN1 bound to the nucleosome.

Clinical immunity to malaria involves epigenetic reprogramming of innate immune cells.

The regulation of inflammation is a critical aspect of disease tolerance and naturally acquired clinical immunity to malaria. Here, we demonstrate using RNA sequencing and epigenetic landscape profiling by cytometry by time-of-flight, that the regulation of inflammatory pathways during asymptomatic parasitemia occurs downstream of pathogen sensing-at the epigenetic level. The abundance of certain epigenetic markers (methylation of H3K27 and dimethylation of arginine residues) and decreased prevalence of histone variant H3.3 correlated with suppressed cytokine responses among monocytes of Ugandan children. Such an epigenetic signature was observed across diverse immune cell populations and not only characterized active asymptomatic parasitemia but also correlated with future long-term disease tolerance and clinical immunity when observed in uninfected children. Pseudotime analyses revealed a potential trajectory of epigenetic change that correlated with a child's age and recent parasite exposure and paralleled the acquisition of clinical immunity. Thus, our data support a model whereby exposure to Plasmodium falciparum induces epigenetic changes that regulate excessive inflammation and contribute to naturally acquire clinical immunity to malaria.

Prospects for breast cancer immunotherapy using microRNAs and transposable elements as objects.

One of the directions in treatment of chemoresistant breast cancer (BC) may include new methods of activating the immune response against tumor cells. Clinically used checkpoint inhibition using antibodies to PD-1 and PD-L1 works in some patients, but the lack of biomarkers means number of respondents is low. The possibility of combining this method with chemotherapy is limited by an increased risk of toxic liver damage, development of immune-related pneumonitis, and thyroid dysfunction. This article includes introduction into the clinic of new methods of immunotherapy for BC, among which epigenetic activation of retroelements, double-stranded transcripts of which stimulate the interferon response against the tumor, is promising. For this purpose, inhibitors of DNA methyltransferase*, histone deacetylase* and histone methyltransferase* are used (* subtitles in the main text). Their antitumor effect is also mediated by removal of repressive epigenetic marks from tumor suppressor genes. However, numerous studies have proven the role of retroelements in the carcinogenesis of various malignant neoplasms, including BC. Moreover, endogenous retroviruses HERV-K and LINE1 retrotransposons are planned to be used as diagnostic biomarkers for BC. Therefore, a rational approach to using viral mimicry in antitumor therapy of BC may be the simultaneous suppression of specific retrotransposons (drivers for carcinogenesis) using reverse transcriptase inhibitors and silencing of specific transposons involved in carcinogenesis using complementary microRNAs. To determine possible pathways of influence in this direction, 35 specific transposon-derived microRNAs* changes in BC were identified, which can become guides for targeted therapy of BC.

Impact of childhood maltreatment on aging: a comprehensive Mendelian randomization analysis of multiple age-related biomarkers

BackgroundChildhood maltreatment (CM) is linked to long-term adverse health outcomes, including accelerated biological aging and cognitive decline. This study investigates the relationship between CM and various aging biomarkers: telomere length, facial aging, intrinsic epigenetic age acceleration (IEAA), GrimAge, HannumAge, PhenoAge, frailty index, and cognitive performance.MethodsWe conducted a Mendelian randomization (MR) study using published GWAS summary statistics. Aging biomarkers included telomere length (qPCR), facial aging (subjective evaluation), and epigenetic age markers (HannumAge, IEAA, GrimAge, PhenoAge). The frailty index was calculated from clinical assessments, and cognitive performance was evaluated with standardized tests. Analyses included Inverse-Variance Weighted (IVW), MR Egger, and Weighted Median (WM) methods, adjusted for multiple comparisons.ResultsCM was significantly associated with shorter telomere length (IVW: β = − 0.1, 95% CI − 0.18 to − 0.02, pFDR = 0.032) and increased HannumAge (IVW: β = 1.33, 95% CI 0.36 to 2.3, pFDR = 0.028), GrimAge (IVW: β = 1.19, 95% CI 0.19 to 2.2, pFDR = 0.040), and PhenoAge (IVW: β = 1.4, 95% CI 0.12 to 2.68, pFDR = 0.053). A significant association was also found with the frailty index (IVW: β = 0.31, 95% CI 0.13 to 0.49, pFDR = 0.006). No significant associations were found with facial aging, IEAA, or cognitive performance.ConclusionsCM is linked to accelerated biological aging, shown by shorter telomere length and increased epigenetic aging markers. CM was also associated with increased frailty, highlighting the need for early interventions to mitigate long-term effects. Further research should explore mechanisms and prevention strategies.

Metaboepigenetic regulation of gene expression in obesity and insulin resistance.

Variation in DNA methylation (DNAm) in adipose tissue is associated with the pathogenesis of obesity and insulin resistance. The activity of enzymes involved in altering DNAm levels is dependent on several metabolite cofactors. To understand the role of metabolites as mechanistic regulators of epigenetic marks, we tested the association between selected plasma metabolites and DNAm levels in the adipose tissue of African Americans. In the AAGMEx cohort (N = 256), plasma levels of metabolites were measured by untargeted liquid chromatography-mass spectrometry; adipose tissue DNAm and transcript levels were measured by reduced representation bisulfite sequencing, and expression microarray, respectively. Among the 21 one-carbon metabolism pathway metabolites evaluated, six were associated with gluco-metabolic traits (PFDR < 0.05, for BMI, SI, or Matsuda index) in AAGMEx. Methylation levels of 196, 116, and 180 CpG-sites were associated (P < 0.0001) with S-adenosylhomocysteine (SAH), cystine, and hypotaurine, respectively. Cis-expression quantitative trait methylation (cis eQTM) analyses suggested the role of metabolite-level-associated CpG sites in regulating the expression of adipose tissue transcripts, including genes in G-protein coupled receptor signaling pathway. Plasma SAH level-associated CpG sites chr19:3403712 and chr19:3403735 were also associated with the expression of G-protein subunit alpha 15 (GNA15) in adipose. The expression of GNA15 was significantly correlated with BMI (β = 1.87, P = 1.9 × 10-16) and SI (β = -1.61, P = 2.49 × 10-5). Our study suggests that a subset of metabolites modulates the methylation levels of CpG sites in specific loci and, in turn, regulates the expression of transcripts involved in obesity and insulin resistance.

Chrom-seq identifies RNAs at chromatin marks.

Chromatin marks are associated with transcriptional regulatory activities. However, very few lncRNAs have been characterized with the role in regulating epigenetic marks, largely due to the technical difficulty in identifying chromatin-associating RNA. Current methods are largely limited by the availability of ChIP-grade antibody and the crosslinking, which generates high noise. Here, we developed a method termed Chrom-seq to efficiently capture RNAs associated with various chromatin marks in living cells. Chrom-seq jointly applies highly specific chromatin mark reader with APEX2, which catalyzes the oxidation of biotin-aniline to label the adjacent RNAs for isolation by streptavidin-coated beads. Using the readers of mCBX7/dPC, mCBX1, and mTAF3, we detected RNA species significantly associated with H3K27me3, H3K9me3, and H3K4me3, respectively. We demonstrated that Chrom-seq outperformed other equivalent methods in terms of sensitivity, efficiency, and cost of practice. It provides an antibody-free approach to systematically map RNAs at chromatin marks with potential regulatory roles in epigenetic events.

Abstract We084: Deciphering the Molecular Code Underlining Postnatal Cardiomyocyte Maturation by an RNA Splicing Factor RBFox1

Mammalian hearts undergo major changes after birth during the perinatal period. While several extrinsic factors such as mechanical load, electrical stimulation, hormones, and nutrients have been implicated in this process, the intrinsic regulatory circuit governing cardiomyocyte postnatal maturation is largely unknown. Importantly, stem-cell derived cardiomyocytes commonly exhibit immaturity which is a major limiting factor for their application as disease models or cell therapy agents. Therefore, uncovering the intrinsic regulatory mechanism of mammalian cardiomyocyte maturation has significant implications in both basic cardiac biology and cell-based therapy and disease modeling. In our recent analysis of the global transcriptome transition from neonatal to adult rat hearts, we found RNA splicing among the top changed pathways. Additionally, we found that expression of RBFox1, a cardiac enriched RNA splicing regulator, was significantly increased during the postnatal transition to adolescence. We further demonstrate that RBFox1 mediated post-transcriptional regulation has a potent effect to promote neonatal and stem-cell derived cardiomyocyte maturation. RNA-seq followed by Gene Ontology enrichment analysis showed that Rbfox1 led to differentially expressed genes enriched in cardiac contraction and conduction, sarcomere organization, K + ion import and muscle filament sliding, as well as differential isoform switches enriched in cardiac contraction and sarcomere structure. An evolutionarily conserved, temporal specific super enhancer exists upstream of cardiac Rbfox1 , which is occupied by active epigenetic markers in the postnatal period. We have identified the tissue and temporal transcriptional activity of this super enhancer in vitro and in vivo . Several binding motifs of transcription factors are predicted within it and a key DNA fragment was identified to be essential for enhancer activity. A transgenic mouse strain has been established to facilitate the tracing of myocyte maturation in vivo by constructing a reporter system driven by a 4xSuperEnhancer-Rbfox1Promoter-tdTomato construct. In conclusion, we propose a mechanism that RBFox1 is an intracellular regulator with tissue- and temporal-specific expression patterns that regulates cardiac maturation by promoting alternative splicing that fine tunes the genetic code.

Current genomic deep learning models display decreased performance in cell type-specific accessible regions

BackgroundA number of deep learning models have been developed to predict epigenetic features such as chromatin accessibility from DNA sequence. Model evaluations commonly report performance genome-wide; however, cis regulatory elements (CREs), which play critical roles in gene regulation, make up only a small fraction of the genome. Furthermore, cell type-specific CREs contain a large proportion of complex disease heritability.ResultsWe evaluate genomic deep learning models in chromatin accessibility regions with varying degrees of cell type specificity. We assess two modeling directions in the field: general purpose models trained across thousands of outputs (cell types and epigenetic marks) and models tailored to specific tissues and tasks. We find that the accuracy of genomic deep learning models, including two state-of-the-art general purpose models―Enformer and Sei―varies across the genome and is reduced in cell type-specific accessible regions. Using accessibility models trained on cell types from specific tissues, we find that increasing model capacity to learn cell type-specific regulatory syntax―through single-task learning or high capacity multi-task models―can improve performance in cell type-specific accessible regions. We also observe that improving reference sequence predictions does not consistently improve variant effect predictions, indicating that novel strategies are needed to improve performance on variants.ConclusionsOur results provide a new perspective on the performance of genomic deep learning models, showing that performance varies across the genome and is particularly reduced in cell type-specific accessible regions. We also identify strategies to maximize performance in cell type-specific accessible regions.

Integrating Multi-Omics with environmental data for precision health: A novel analytic framework and case study on prenatal mercury induced childhood fatty liver disease

BackgroundPrecision Health aims to revolutionize disease prevention by leveraging information across multiple omic datasets (multi-omics). However, existing methods generally do not consider personalized environmental risk factors (e.g., environmental pollutants). ObjectiveTo develop and apply a precision health framework which combines multiomic integration (including early, intermediate, and late integration, representing sequential stages at which omics layers are combined for modeling) with mediation approaches (including high-dimensional mediation to identify biomarkers, mediation with latent factors to identify pathways, and integrated/quasi-mediation to identify high-risk subpopulations) to identify novel biomarkers of prenatal mercury induced metabolic dysfunction-associated fatty liver disease (MAFLD), elucidate molecular pathways linking prenatal mercury with MAFLD in children, and identify high-risk children based on integrated exposure and multiomics data. MethodsThis prospective cohort study used data from 420 mother–child pairs from the Human Early Life Exposome (HELIX) project. Mercury concentrations were determined in maternal or cord blood from pregnancy. Cytokeratin 18 (CK-18; a MAFLD biomarker) and five omics layers (DNA Methylation, gene transcription, microRNA, proteins, and metabolites) were measured in blood in childhood (age 6–10 years). ResultsEach standard deviation increase in prenatal mercury was associated with a 0.11 [95% confidence interval: 0.02–0.21] standard deviation increase in CK-18. High dimensional mediation analysis identified 10 biomarkers linking prenatal mercury and CK-18, including six CpG sites and four transcripts. Mediation with latent factors identified molecular pathways linking mercury and MAFLD, including altered cytokine signaling and hepatic stellate cell activation. Integrated/quasi-mediation identified high risk subgroups of children based on unique combinations of exposure levels, omics profiles (driven by epigenetic markers), and MAFLD. ConclusionsPrenatal mercury exposure is associated with elevated liver enzymes in childhood, likely through alterations in DNA methylation and gene expression. Our analytic framework can be applied across many different fields and serve as a resource to help guide future precision health investigations.

Epigenetic disease markers in primary sclerosing cholangitis and primary biliary cholangitis-methylomics of cholestatic liver disease.

The epigenome, the set of modifications to DNA and associated molecules that control gene expression, cellular identity, and function, plays a major role in mediating cellular responses to outside factors. Thus, evaluation of the epigenetic state can provide insights into cellular adaptions occurring over the course of disease. We performed epigenome-wide association studies of primary sclerosing cholangitis (PSC) and primary biliary cholangitis (PBC) using the Illumina MethylationEPIC Bead Chip. We found evidence of increased epigenetic age acceleration and differences in predicted immune cell composition in patients with PSC and PBC. Epigenetic profiles demonstrated differences in predicted protein levels including increased levels of tumor necrosis factor receptor superfamily member 1B in patients with cirrhotic compared to noncirrhotic PSC and PBC. Epigenome-wide association studies of PSC discovered strongly associated 5'-C-phosphate-G-3' sites in genes including vacuole membrane protein 1 and SOCS3, and epigenome-wide association studies of PBC found strong 5'-C-phosphate-G-3' associations in genes including NOD-like receptor family CARD domain containing 5, human leukocyte antigen-E, and PSMB8. Analyses identified disease-associated canonical pathways and upstream regulators involved with immune signaling and activation of macrophages and T-cells. A comparison of PSC and PBC data found relatively little overlap at the 5'-C-phosphate-G-3' and gene levels with slightly more overlap at the level of pathways and upstream regulators. This study provides insights into methylation profiles of patients that support current concepts of disease mechanisms and provide novel data to inspire future research. Studies to corroborate our findings and expand into other -omics layers will be invaluable to further our understanding of these rare diseases with the goal to improve and individualize prognosis and treatment.