DNA Methylation Microarray Data Research Articles

Immunological shifts during early-stage Parkinson’s disease identified with DNA methylation data on longitudinally collected blood samples

Parkinson’s disease (PD) is the second most common neurodegenerative disease in the United States. Decades before motor symptoms manifest, non-motor symptoms such as hyposmia and rapid eye movement (REM) sleep behavior disorder are highly predictive of PD. Previous immune profiling studies have identified alterations to the proportions of immune cells in the blood of clinically defined PD patients. However, it remains unclear if these phenotypes manifest before the clinical diagnosis of PD. We utilized longitudinal DNA methylation (DNAm) microarray data from the Parkinson’s Progression Marker’s Initiative (PPMI) to perform immune profiling in clinically defined PD and prodromal PD patients (Prod). We identified previously reported changes in neutrophil, monocyte, and T cell numbers in PD patients. Additionally, we noted previously unrecognized decreases in the naive B cell compartment in the defined PD and Prod patient group. Over time, we observed the proportion of innate immune cells in PD blood increased, but the proportion of adaptive immune cells decreased. We identified decreases in T and B cell subsets associated with REM sleep disturbances and early cognitive decline. Lastly, we identified increases in B memory cells associated with both genetic (LRRK2 genotype) and infectious (cytomegalovirus seropositivity) risk factors of PD. Our analysis shows that the peripheral immune system is dynamic as the disease progresses. The study provides a platform to understand how and when peripheral immune alterations occur in PD and whether intervention at particular stages may be therapeutically advantageous.

Simultaneous assessment of mitochondrial DNA copy number and nuclear epigenetic age towards predictive models of development and aging

ObjectiveMitochondrial dysfunction and nuclear epigenetic alterations, two hallmarks of aging, are associated with aberrant development and complex disease risk. Here, we report a method for the simultaneous assessment of mitochondrial DNA copy number (mtDNA-CN) and DNA methylation age (DNAm age) from the same DNA extraction using quantitative polymerase chain reaction (qPCR) and array data, respectively.ResultWe present methods for the concurrent estimation of mtDNA-CN and DNAm age from the same DNA samples. This includes qPCR to estimate mtDNA-CN, representing the number of circular mitochondrial genomes in a cell, and DNA methylation microarray data to estimate the epigenetic age of an individual. Further, we provide a method for the combination of these metrics into a shared metric termed ‘mtEpiAge’. This approach provides a valuable tool for exploring the interplay between mitochondrial dysfunction and nuclear epigenetic alterations, and their associations with disease and aging.

Genome-wide DNA methylation analysis identifies potent CpG signature for temzolomide response in non-G-CIMP glioblastomas with unmethylated MGMT promoter: MGMT-dependent roles of GPR81.

To identify potent DNA methylation candidates that could predict response to temozolomide (TMZ) in glioblastomas (GBMs) that do not have glioma-CpGs island methylator phenotype (G-CIMP) but have an unmethylated promoter of O-6-methylguanine-DNA methyltransferase (unMGMT). The discovery-validation approach was planned incorporating a series of G-CIMP-/unMGMT GBM cohorts with DNA methylation microarray data and clinical information, to construct multi-CpG prediction models. Different bioinformatic and experimental analyses were performed for biological exploration. By analyzing discovery sets with radiotherapy (RT) plus TMZ versus RT alone, we identified a panel of 64 TMZ efficacy-related CpGs, from which a 10-CpG risk signature was further constructed. Both the 64-CpG panel and the 10-CpG risk signature were validated showing significant correlations with overall survival of G-CIMP-/unMGMT GBMs when treated with RT/TMZ, rather than RT alone. The 10-CpG risk signature was further observed for aiding TMZ choice by distinguishing differential outcomes to RT/TMZ versus RT within each risk subgroup. Functional studies on GPR81, the gene harboring one of the 10 CpGs, indicated its distinct impacts on TMZ resistance in GBM cells, which may be dependent on the status of MGMT expression. The 64 TMZ efficacy-related CpGs and in particular the 10-CpG risk signature may serve as promising predictive biomarker candidates for guiding optimal usage of TMZ in G-CIMP-/unMGMT GBMs.

Circulating tumor DNA (ctDNA) methylation-based linkage regions to predict recurrence in early-stage lung cancer.

e20513 Background: Lung cancer is the most incident worldwide, and surveillance of recurrence remains a clinically unmet need. Non-invasive early detection is essential to improve prognosis of lung cancer. DNA methylation-based biomarkers for early cancer detection is promising. This study aims to investigate the methylation biomarkers of ctDNA that could predict the postoperative recurrence in early-stage lung cancer. Methods: The HM450K DNA methylation microarray data of lung adenocarcinoma (LUAD) (492 tumor tissues and 32 normal tissues) and lung squamous cell carcinoma (LUSC) (415 tumor tissues and 43 normal tissues) were downloaded from TCGA Data Portal ( https://portal.gdc.cancer.gov/ ), and healthy individuals were obtained from GEO database (656 whole blood from GSE40279) ( https://www.ncbi.nlm.nih.gov/geo/ ). After processing missing values with impute.knn function in CHAMP package, differences > 0.1 between tumor tissues and control samples (normal tissues or whole blood), standard deviation (SD) < 0.15 for tumor tissues and SD < 0.1 for control samples were nominated as differentially methylated sites (DMS). The adjacent ( < 250 bp) methylation sites were stitched into peaks. MCBs (methylation-correlated blocks) were defined as regions with more than or equal to 3 CpG sites within 100 bp and with DNA methylation sites correlation > 0.5. Results: In TCGA LUAD and LUSC, 658 and 551 DMS were identified between tumor and normal tissues; 999 and 912 DMS were generated by comparing tumor tissues and whole blood. The intersection of DMS located in the methylation canyon created a 203 kbp DNA methylation panel, containing 2521 DMS and 1417 methylation peaks. 37 MCBs were detected in tumor tissues of 30 lung cancer with DNA methylation panel sequencing. In term of 37 MCBs, 4 MCBs located in SHOX2, PTGER4, RASSF1A and ARL8B were identified with significant differences (p < 0.001) between plasma of 50 lung cancer patients and 50 healthy individuals. Furthermore, Real-time Quantitative PCR (qPCR) of the 4 MCB regions were performed in plasma of 30 lung cancer patients and 25 healthy individuals, and a cutoff of 0.414 distinguishing lung cancer from the healthy was calculated by the generalized linear model (GLM) based on the cycle threshold (CT) values of qPCR. The GLM incorporated the 4 MCBs showed robust performances in which the sensitivity, specificity and area under the curve (AUC) were 100%, 95% and 0.95, respectively. Postoperative plasma within 3 months were retrospectively collected from 32 lung cancer patients, and 30 patients with GLM scored less than 0.414 (Maximum: 0.327) had no recurrence within 1 year, while 2 patients with GLM scored more than 0.414 (0.480 and 0.557) were recurred within 1 year. Conclusions: The generalized linear model incorporated the 4 MCBs showed robust performances, and predicted the recurrence of early-stage lung cancer within 1 year after surgery.

Abstract LB058: A transcriptome-wide gene expression outlier analysis pinpoints therapeutic vulnerabilities in colorectal cancer

Abstract The number of effective therapies approved for metastatic colorectal cancer (mCRC) is still limited and multiple strategies are continuously explored to expand the drug target repertoire. Among these, the identification of overexpressed genes has prompted the discovery of actionable oncogenic dependencies in multiple tumour types. Starting from RNA sequencing data, we identified transcriptome-wide gene expression outliers, defined as samples showing abnormal expression for a particular gene, across 226 CRC cell lines, considering both overexpression and underexpression events as positive or negative outliers. Then, the distance of each outlier from gene-specific reference points, absolute expression values and differential expression values were considered in a multi-filter strategy to select extreme gene expression outliers, with the hypothesis that they are more likely to be functionally relevant in cancer cells. We also profiled genetic and epigenetic features of CRC cell lines based on whole exome sequencing and DNA methylation microarray data. Extreme positive and negative gene expression outliers were found for 3,533 and 965 genes, respectively, and only some of them were associated with underlying genetic and epigenetic alterations. Gene expression alterations with known therapeutic or diagnostic value in CRC were pinpointed as extreme positive and negative outliers thus confirming the validity of the approach. Annotation of overexpressed enzyme genes according to the Target Development Level (TDL) classification revealed numerous enzymes for which inhibitors are already available. We next explored underexpression events to identify potential synthetic lethal targets. Intriguingly, we found that CRC models lacking expression of the MTAP gene were sensitive to treatment with an inhibitor of the PRMT5:MTA complex currently under clinical development. We found that mapping extreme and transcriptome-wide positive and negative gene expression outliers in CRC cell lines is an effective strategy to identify putative drug targets and biomarkers, independently from the underlying genetic or epigenetic alterations. We indeed present a comprehensive atlas of CRC extreme gene expression outliers which includes events with diagnostic or therapeutic relevance. This resource could also serve as a reference for further discoveries in CRC and other tumour types. Citation Format: Gaia Grasso, Elisa Marriella, Martina Miotto, Kristi Buzo, Nicole M. Reilly, Pietro Andrei, Pietro P. Vitiello, Giovanni Crisafulli, Sabrina Arena, Giuseppe Rospo, Giorgio Corti, Annalisa Lorenzato, Carlotta Cancelliere, Ludovic Barault, Giulia Gionfriddo, Michael Linnebacher, Mariangela Russo, Federica Di Nicolantonio, Alberto Bardelli. A transcriptome-wide gene expression outlier analysis pinpoints therapeutic vulnerabilities in colorectal cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 2 (Clinical Trials and Late-Breaking Research); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(8_Suppl):Abstract nr LB058.

Deep belief network-based approach for detecting Alzheimer's disease using the multi-omics data

Alzheimer's disease (AD) is the most uncertain form of Dementia in terms of finding out the mechanism. AD does not have a vital genetic factor to relate to. There were no reliable techniques and methods to identify the genetic risk factors associated with AD in the past. Most of the data available were from the brain images. However, recently, there have been drastic advancements in the high-throughput techniques in bioinformatics. It has led to focused researches in discovering the AD causing genetic risk factors. Recent analysis has resulted in considerable prefrontal cortex data with which classification and prediction models can be developed for AD. We have developed a Deep Belief Network-based prediction model using the DNA Methylation and Gene Expression Microarray Data, with High Dimension Low Sample Size (HDLSS) issues. To overcome the HDLSS challenge, we performed a two-layer feature selection considering the biological aspects of the features as well. In the two-layered feature selection approach, first the differentially expressed genes and differentially methylated positions are identified, then both the datasets are combined using Jaccard similarity measure. As the second step, an ensemble-based feature selection approach is implemented to further narrow down the gene selection. The results show that the proposed feature selection technique outperforms the existing commonly used feature selection techniques, such as Support Vector Machine Recursive Feature Elimination (SVM-RFE), and Correlation-based Feature Selection (CBS). Furthermore, the Deep Belief Network-based prediction model performs better than the widely used Machine Learning models. Also, the multi-omics dataset shows promising results compared to the single omics.

Elevated BICD2 DNA methylation in blood of major depressive disorder patients and reduction of depressive-like behaviors in hippocampal Bicd2-knockdown mice

Major depressive disorder (MDD) is a prevalent and devastating mental illness. To date, the diagnosis of MDD is largely dependent on clinical interviews and questionnaires and still lacks a reliable biomarker. DNA methylation has a stable and reversible nature and is likely associated with the course and therapeutic efficacy of complex diseases, which may play an important role in the etiology of a disease. Here, we identified and validated a DNA methylation biomarker for MDD from four independent cohorts of the Chinese Han population. First, we integrated the analysis of the DNA methylation microarray (n = 80) and RNA expression microarray data (n = 40) and identified BICD2 as the top-ranked gene. In the replication phase, we employed the Sequenom MassARRAY method to confirm the DNA hypermethylation change in a large sample size (n = 1,346) and used the methylation-sensitive restriction enzymes and a quantitative PCR approach (MSE-qPCR) and qPCR method to confirm the correlation between DNA hypermethylation and mRNA down-regulation of BICD2 (n = 60). The results were replicated in the peripheral blood of mice with depressive-like behaviors, while in the hippocampus of mice, Bicd2 showed DNA hypomethylation and mRNA/protein up-regulation. Hippocampal Bicd2 knockdown demonstrates antidepressant action in the chronic unpredictable mild stress (CUMS) mouse model of depression, which may be mediated by increased BDNF expression. Our study identified a potential DNA methylation biomarker and investigated its functional implications, which could be exploited to improve the diagnosis and treatment of MDD.

InterpolatedXY: a two-step strategy to normalize DNA methylation microarray data avoiding sex bias.

MotivationData normalization is an essential step to reduce technical variation within and between arrays. Due to the different karyotypes and the effects of X chromosome inactivation, females and males exhibit distinct methylation patterns on sex chromosomes; thus, it poses a significant challenge to normalize sex chromosome data without introducing bias. Currently, existing methods do not provide unbiased solutions to normalize sex chromosome data, usually, they just process autosomal and sex chromosomes indiscriminately.ResultsHere, we demonstrate that ignoring this sex difference will lead to introducing artificial sex bias, especially for thousands of autosomal CpGs. We present a novel two-step strategy (interpolatedXY) to address this issue, which is applicable to all quantile-based normalization methods. By this new strategy, the autosomal CpGs are first normalized independently by conventional methods, such as funnorm or dasen; then the corrected methylation values of sex chromosome-linked CpGs are estimated as the weighted average of their nearest neighbors on autosomes. The proposed two-step strategy can also be applied to other non-quantile-based normalization methods, as well as other array-based data types. Moreover, we propose a useful concept: the sex explained fraction of variance, to quantitatively measure the normalization effect.Availability and implementationThe proposed methods are available by calling the function ‘adjustedDasen’ or ‘adjustedFunnorm’ in the latest wateRmelon package (https://github.com/schalkwyk/wateRmelon), with methods compatible with all the major workflows, including minfi.Supplementary information Supplementary data are available at Bioinformatics online.

A Novel Pseudogene Methylation Signature to Predict Temozolomide Outcome in Non-G-CIMP Glioblastomas.

Objective Alterations in the methylation state of pseudogenes may serve as clinically useful biomarkers of glioblastomas (GBMs) that do not have glioma-CpG island methylator phenotype (G-CIMP). Methods Non-G-CIMP GBM datasets were included for evaluation, and a RISK-score signature was determined from the methylation state of pseudogene loci. Both bioinformatic and experimental analyses were performed for biological validation. Results By integrating clinical information with DNA methylation microarray data, we screened a panel of eight CpGs from discovery cohorts of non-G-CIMP GBMs. Each CpG could accurately and independently predict the prognosis of patients under a treatment regime that combined radiotherapy (RT) and temozolomide (TMZ). The 8-CpG signature appeared to show opposite prognostic correlations between patients treated with RT/TMZ and those treated with RT monotherapy. The analyses further indicated that this signature had predictive value for TMZ efficacy because different survival benefits between RT/TMZ and RT therapies were observed in each risk subgroup. The incorporation of other risk factors, such as age and O-6-methylguanine-DNA methyltransferase (MGMT) promoter methylation status, with our pseudogene methylation signature could provide precise risk classification. In vitro experimental data revealed that two locus-specific pseudogenes (ZNF767P and CLEC4GP1) may modulate TMZ resistance via distinct mechanisms in GBM cells. Conclusion The biologically and clinically relevant RISK-score signature, based on pseudogene methylation loci, may offer information for predicting TMZ responses of non-G-CIMP GBMs, that is independent from, but complementary to, MGMT-based approaches.

PATH-06. Molecular subgrouping of medulloblastoma via low-depth whole genome bisulfite sequencing

INTRODUCTION: International consensus recognises four molecular subgroups of medulloblastoma, each with distinct molecular features and clinical outcomes. Assigning molecular subgroup is typically achieved via the Illumina DNA methylation microarray. Given the rapidly-expanding WGS capacity in healthcare institutions, there is an unmet need to develop platform-independent, sequence-based subgrouping assays. Whole genome bisulfite sequencing (WGBS) enables the assessment of genome-wide methylation status at single-base resolution. To date, its routine application for subgroup assignment has been limited, due to high economic cost and sample input requirements. Currently, no optimised pipeline exists that is tailored to handle samples sequenced at low-pass (i.e., <10x depth). METHODOLOGY: Two datasets were utilised; 36 newly sequenced low-depth (10x) and 42 publicly available high-depth (30x) WGBS medulloblastoma samples (n=34), alongside cerebellar control samples (n=8), all with matched DNA methylation microarray data. We applied imputation to low-pass WGBS data, assessed inter-platform correlation and identified molecular subgroups by directly integrating WGBS sample data with pre-existing array-trained models. We developed machine learning WGBS-based classifiers and compared performance against microarray. We optimised reference-free aneuploidy detection with low-pass WGBS and assessed concordance with microarray-derived aneuploidy calls. RESULTS: We optimised a pipeline for processing, QC, and analysis of low-pass WGBS data, suitable for routine molecular subgrouping and reference-free aneuploidy assessment that achieves 96% sensitivity compared to microarray approaches. A pilot study of the suitability of FFPE was promising, and we demonstrate that WGBS data can be integrated into existing array-trained models with high assignment probabilities. Also, WGBS-derived classifier performance measures exceeded microarray-derived classifiers. CONCLUSION: We describe a platform-independent WGBS assay for molecular subgrouping of medulloblastoma. It performs equivalently to array-based methods at increasingly comparable cost ($400 vs $580) and provides a proof-of-concept for routine clinical adoption using standard WGS technology. Finally, the full methylome enabled elucidation of additional biological heterogeneity that has hitherto been inaccessible.

DNA methylation profiling to determine the primary sites of metastatic cancers using formalin-fixed paraffin-embedded tissues.

3079 Background: Metastatic cancers with uncertain primary sites account for a significant portion of new cases. Among them, 3-9% are eventually assigned to cancer of unknown primary (CUP) site after a comprehensive diagnostic workup. Accurate identification of the primary site is the starting point for cancer diagnosis worldwide, and it is critical to guide the subsequent treatments of metastatic cancers. Here we presented a new DNA methylation sequencing-based method to predict the tissues of origin for metastatic cancers, including CUP. Methods: Cancer diagnosis relies substantially on histological and immunohistochemical analyses of formalin-fixed paraffin-embedded (FFPE) tissues. To take advantage of sample accessibility, we developed an optimized and streamlined method that was particularly used to generate reduced represent bisulfite sequencing (RRBS) libraries for genome-wide DNA methylation profiling with degraded DNA fragments. After confirming that data quality generated using the new FFPE-RRBS method was comparable with regular RRBS, we created an RRBS database using 541 fresh frozen samples across ten most common cancer types and 58 tumor-adjacent normal tissues. By incorporating four distinct methylation summary scores and seven machine learning approaches, 28 models were trained and compared for multi-class classification using our database. Lastly, we selected the best classifier to predict the tissues of origin utilizing 249 FFPE samples across ten metastatic cancer types and 12 FFPE samples from CUP patients. Meanwhile, the classifier was also cross-validated using a DNA methylation microarray data set of 4702 patients diagnosed with corresponding primary cancers in the TCGA project. Results: FFPE-RRBS allowed to construct decent libraries with heavily degraded genomic DNA within 20 hours. Comparable DNA methylation metrics were obtained for the RRBS libraries of paired primary cancer tissues (fresh frozen vs. FFPE) and the libraries of paired FFPE samples derived from primary and metastatic tissues. Among the 28 methylation-based classifiers, the mean methylation-based LinearSVC model performed the best, achieving an overall accuracy of 81% with an AUC of 0.95 in determining the primary sites of 10 metastatic cancers. In a cross-validation assay using the TCGA data set of 4702 cancer patients, the overall prediction accuracy and AUC were 92% and 0.99, respectively. Lastly, our model successfully identified the tissues of origin in 10 of 12 CUP patients in a prospective study. Conclusions: The FFPE-RRBS is a novel method for efficient profiling of heavily degraded FFPE samples and the mean methylation-based LinearSVC model can predict the tissues of origin for metastatic cancers and CUP with high accuracy.

DNA Methylation Patterns According to Fatty Liver Index and Longitudinal Changes from the Korean Genome and Epidemiology Study (KoGES).

The role of differentially methylated regions (DMRs) in nonalcoholic fatty liver disease (NAFLD) is unclear. This study aimed to identify the role of DMR in NAFLD development and progression using the Korean Genome and Epidemiology Study (KoGES) cohort. We used laboratory evaluations and Illumina Methylation 450 k DNA methylation microarray data from KoGES. The correlation between fatty liver index (FLI) and genomic CpG sites was analyzed in 322 subjects. Longitudinal changes over 8 years were confirmed in 33 subjects. To identify CpG sites and genes related to FLI, we obtained enrichment terms for 6765 genes. DMRs were identified for both high (n = 128) and low (n = 194) groups on the basis of FLI 30 in 142 men and 180 women. To confirm longitudinal changes in 33 subjects, the ratio of follow-up and baseline investigation values was obtained. Correlations and group comparisons were performed for the 8 year change values. PITPNM3, RXFP3, and THRB were hypermethylated in the increased FLI groups, whereas SLC9A2 and FOXI3 were hypermethylated in the decreased FLI groups. DMRs describing NAFLD were determined, and functions related to inflammation were identified. Factors related to longitudinal changes are suggested, and blood circulation-related functions appear to be important in the management of NAFLD.

Identification potential epigenetic biomarkers of a human immunodeficiency virus/tuberculosis co-infection based on weighted gene co-expression network analysis.

Tuberculosis (TB) is one of the most common opportunistic infections and a leading cause of death in patients infected with human immunodeficiency virus (HIV). However, conventional diagnostic tools have several limitations. The aim of this study was to screen key DNA methylated cytosine-phosphate-guanine dinucleotide (CpG) islands (CGIs) to identify potential diagnosis biomarkers in HIV mono-infected patients and HIV/TB co-infected patients based on a network analysis. The GSE50835 DNA methylation microarray data were downloaded from the Gene Expression Omnibus (GEO) database. Differentially methylated CpG islands analysis, weighted gene co-expression network analysis (WGCNA), and least absolute shrinkage and selection operator (LASSO) logistic regression were performed in 19 HIV mono-infected patients and 20 HIV/TB co-infected patients. In total, 1950 differentially methylated CpG islands were identified, and weighted co-methylation network construction and module preservation revealed one network module that can distinguish the HIV/TB co-infected patients from the HIV mono-infected patients. Based on the LASSO logistic regression, an eight-methylated CpG island diagnosis model was established that can accurately distinguish HIV/TB co-infected patients from HIV mono-infected patients with a sensitivity of 87.2%, a specificity of 88.7%, and an area under the receiver operating characteristic (ROC) curve (AUC) of 0.948. Alteration in the eight-DNA methylated CpG sites might be involved in the pathology of an HIV/TB co-infection and could be used as potential diagnosis biomarkers in HIV/TB co-infected patients.

DNA Methylation Profiling for the Diagnosis and Prognosis of Patients with Nontuberculous Mycobacterium Lung Disease

The incidence of nontuberculous Mycobacterium (NTM) lung disease is rapidly increasing; however, its diagnosis and prognosis remain unclear while selecting patients who will respond to appropriate treatment. Differences in DNA methylation patterns between NTM patients with good or poor prognosis could provide important therapeutic targets. We used the Illumina MethylationEPIC (850k) DNA methylation microarray to determine the pattern between differentially methylated regions (DMRs) in NTM patients with good or poor prognosis (n = 4/group). Moreover, we merged and compared 20 healthy controls from previous Illumina Methylation450k DNA methylation microarray data. We selected and visualized the DMRs in the form of heatmaps, and enriched terms associated with these DMRs were identified by functional annotation with the “pathfinder” package. In total, 461 and 293 DMRs (|Log2 fold change| > 0.1 and P < 0.03) were more methylated in patients with four poor and four good prognoses, respectively. Furthermore, 337 and 771 DMRs (|Log2 fold change| > 0.08 and P < 0.001) were more methylated in eight NTM patients and 20 healthy controls, respectively. TGFBr1 was significantly less methylated, whereas HLA-DR1 and HLA-DR5 were more methylated in patients with poor prognosis (compared to those with good prognosis). LRP5, E2F1, and ADCY3 were the top three less-methylated genes in NTM patients (compared with the controls). The mTOR and Wnt signaling pathway-related genes were less methylated in patients with NTM. Collectively, genes related to Th1- cell differentiation, such as TGFBr1 and HLA-DR, may be used as biomarkers for predicting the treatment response in patients with NTM lung disease.

OMIC-06. MOLECULAR SUBGROUPING OF MEDULLOBLASTOMA VIA LOW-DEPTH WHOLE GENOME BISULFITE SEQUENCING

IntroductionInternational consensus recognises four molecular subgroups of medulloblastoma, each with distinct molecular features and clinical outcomes. The current gold-standard for subgroup assignment is DNA methylation microarray. There is an unmet need to develop platform-independent subgrouping assays which are both non-proprietary and compatible with rapidly-expanding WGS capacity in healthcare. Whole Genome Bisulfite Sequencing (WGBS) enables the assessment of genome-wide methylation status at single-base resolution. Previously, WGBS adoption has been limited by cost and sample quality/quantity requirements. Its application for routine detection of medulloblastoma subgroups has not previously been reported. MethodologyTwo datasets were utilised; 36 newly-sequenced low-depth (10x coverage) and 34 publicly-available high-depth (30x) WGBS medulloblastomas, all with matched DNA methylation microarray data. We compared platform concordance and identified molecular subgroups. Machine-learning WGBS-based subgroup classifiers were optimised and compared between platforms. Aneuploidy and mutation detection using WGBS was optimised and compared to microarray-derived estimates where possible. Finally, comprehensive subgroup-specific DNA methylation signatures were identified. ResultsWe optimised a pipeline for processing, quality control and analysis of low-depth WGBS data, suitable for routine molecular subgrouping and aneuploidy assessment. We demonstrated the suitability of fresh-frozen and FFPE DNA for WGBS, and, using downsampling, showed that subgroup calling is robust at coverages as low as 2x. We identified differentially methylated regions that, due to poor representation, could not be detected using methylation microarrays. Molecular subgroups of medulloblastoma assigned using WGBS were concordant with array-based definitions, and WGBS-derived classifier performance measures exceeded microarray-derived classifiers. ConclusionWe describe a platform-independent assay for molecular subgrouping of medulloblastoma using WGBS. It performs equivalently to current array-based methods at comparable cost ($405 vs $596) and provides a proof-of-concept for its routine clinical adoption using standard WGS technology. Finally, the full methylome enabled elucidation of additional biological heterogeneity that has hitherto been inaccessible.

Male-specific age estimation based on Y-chromosomal DNA methylation.

Although DNA methylation variation of autosomal CpGs provides robust age predictive biomarkers, no male-specific age predictor exists based on Y-CpGs yet. Since sex chromosomes play an important role in aging, a Y-chromosome-based age predictor would allow studying male-specific aging effects and would also be useful in forensics. Here, we used blood-based DNA methylation microarray data of 1,057 males from six cohorts aged 15-87 and identified 75 Y-CpGs with an interquartile range of ≥0.1. Of these, 22 and six were significantly hyper- and hypomethylated with age (p(cor)<0.05, Bonferroni), respectively. Amongst several machine learning algorithms, a model based on support vector machines with radial kernel performed best in male-specific age prediction. We achieved a mean absolute deviation (MAD) between true and predicted age of 7.54 years (cor=0.81, validation) when using all 75 Y-CpGs, and a MAD of 8.46 years (cor=0.73, validation) based on the most predictive 19 Y-CpGs. The accuracies of both age predictors did not worsen with increased age, in contrast to autosomal CpG-based age predictors that are known to predict age with reduced accuracy in the elderly. Overall, we introduce the first-of-its-kind male-specific epigenetic age predictor for future applications in aging research and forensics.

Chinese Glioma Genome Atlas (CGGA): A Comprehensive Resource with Functional Genomic Data from Chinese Glioma Patients

Gliomas are the most common and malignant intracranial tumors in adults. Recent studies have revealed the significance of functional genomics for glioma pathophysiological studies and treatments. However, access to comprehensive genomic data and analytical platforms is often limited. Here, we developed the Chinese Glioma Genome Atlas (CGGA), a user-friendly data portal for the storage and interactive exploration of cross-omics data, including nearly 2000 primary and recurrent glioma samples from Chinese cohort. Currently, open access is provided to whole-exome sequencing data (286 samples), mRNA sequencing (1018 samples) and microarray data (301 samples), DNA methylation microarray data (159 samples), and microRNA microarray data (198 samples), and to detailed clinical information (age, gender, chemoradiotherapy status, WHO grade, histological type, critical molecular pathological information, and survival data). In addition, we have developed several tools for users to analyze the mutation profiles, mRNA/microRNA expression, and DNA methylation profiles, and to perform survival and gene correlation analyses of specific glioma subtypes. This database removes the barriers for researchers, providing rapid and convenient access to high‐quality functional genomic data resources for biological studies and clinical applications. CGGA is available at http://www.cgga.org.cn.

Integrative analysis identifies an immune-relevant epigenetic signature for prognostication of non-G-CIMP glioblastomas

ABSTRACT The clinical and molecular implications of DNA methylation alterations remain unclear among the majority of glioblastomas (GBMs) without glioma-CpGs island methylator phenotype (G-CIMP); integrative multi-level molecular profiling may provide useful information. Independent cohorts of non-G-CIMP GBMs or IDH wild type (wt) lower-grade gliomas (LGGs) from local and public databases with DNA methylation and gene expression microarray data were included for discovery and validation of a multimarker signature, combined using a RISK score model. Bioinformatic and in vitro functional analyses were employed for biological validation. Using a strict multistep selection approach, we identified eight CpGs, each of which was significantly correlated with overall survival (OS) of non-G-CIMP GBMs, independent of age, the O-6-methylguanine-DNA methyltransferase (MGMT) methylation status, treatments and other identified CpGs. An epigenetic RISK signature of the 8 CpGs was developed and validated to robustly and independently prognosticate prognosis in different cohorts of not only non-G-GIMP GBMs, but also IDHwt LGGs. It also showed good discriminating value in stratified cohorts by current clinical and molecular factors. Bioinformatic analysis revealed consistent correlation of the epigenetic signature to distinct immune-relevant transcriptional profiles of GBM bulks. Functional experiments showed that S100A2 appeared to be epigenetically regulated by one identified CpG and was associated with GBM cell proliferation, apoptosis, invasion, migration and immunosuppression. The prognostic 8-CpGs RISK score signature may be of promising value for refining current glioma risk classification, and its potential links to distinct immune phenotypes make it a promising biomarker candidate for predicting response to anti-glioma immunotherapy.

Abstract 3661: Association of mammographic density with blood DNA methylation

Abstract Background: Mammographic density is a strong risk factor for breast cancer. While numerous genetic variations have been shown to be associated with mammographic density, an epigenetic signature of mammographic density has not yet been identified. The objective of this study was to identify a DNA methylation signature associated with mammographic density. Methods: Blood samples were collected from 197 postmenopausal women 45-65 years old with no personal history of breast cancer. For each participant, mammographic density according to the BI-RADs density classification was obtained from the most recent mammogram report. White blood cell DNA methylation at more than 850,000 CpG sites was measured using the Illumina HumanMethylationEPIC BeadChip. Differential methylation was assessed using genome-wide, probe-level, and regional approaches after filtering and normalization using the recommended steps for DNA methylation microarray data. All analyses were adjusted for appropriate confounding variables. Results: On average, genome-wide DNA methylation was higher in women with elevated mammographic density. Additionally, DNA methylation at 343 individual sites and in 34 regions was significantly associated with mammographic density after false discovery rate correction. Most density-associated sites were hypermethylated. The top differentially methylated sites were located near cell cycle regulatory genes (ICMT, DESI2, TCF21/TARID, ROBO3), as well as those encoding cytoskeletal or plasma membrane proteins (PALLD, ANO7, ATP8B3, MTUS2). We observed hypermethylation near the promoters of genes including RUFY1 and SLFN12 and within the bodies of genes including LCN6, SNED1, and LRFN1. Functional analysis of the differentially methylated sites revealed enrichment of gene ontology categories including angiogenesis and regulation of cell adhesion, as well as apoptotic signaling in response to DNA damage. Conclusions: White blood cell DNA methylation differed according to mammographic density in a sample of postmenopausal women. Probe-based and regional analysis identified differentially methylated regions located near genes implicated in cancer. These differentially methylated regions may serve as a signature of mammographic density, and future studies of the genes implicated may provide a deeper understanding of the connection between mammographic density and breast cancer risk. Citation Format: Rachel McFarland Lucia, Wei-Lin Huang, Andrea Alvarez, Irene Masunaka, Argyrios Ziogas, Deborah Goodman, Andrew O. Odegaard, Trina M. Norden-Krichmar, Hannah Lui Park. Association of mammographic density with blood DNA methylation [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 3661.

Simulating ComBat: how batch correction can lead to the systematic introduction of false positive results in DNA methylation microarray studies

BackgroundSystematic technical effects—also called batch effects—are a considerable challenge when analyzing DNA methylation (DNAm) microarray data, because they can lead to false results when confounded with the variable of interest. Methods to correct these batch effects are error-prone, as previous findings have shown.ResultsHere, we demonstrate how using the R function ComBat to correct simulated Infinium HumanMethylation450 BeadChip (450 K) and Infinium MethylationEPIC BeadChip Kit (EPIC) DNAm data can lead to a large number of false positive results under certain conditions. We further provide a detailed assessment of the consequences for the highly relevant problem of p-value inflation with subsequent false positive findings after application of the frequently used ComBat method. Using ComBat to correct for batch effects in randomly generated samples produced alarming numbers of false discovery rate (FDR) and Bonferroni-corrected (BF) false positive results in unbalanced as well as in balanced sample distributions in terms of the relation between the outcome of interest variable and the technical position of the sample during the probe measurement. Both sample size and number of batch factors (e.g. number of chips) were systematically simulated to assess the probability of false positive findings. The effect of sample size was simulated using n = 48 up to n = 768 randomly generated samples. Increasing the number of corrected factors led to an exponential increase in the number of false positive signals. Increasing the number of samples reduced, but did not completely prevent, this effect.ConclusionsUsing the approach described, we demonstrate, that using ComBat for batch correction in DNAm data can lead to false positive results under certain conditions and sample distributions. Our results are thus contrary to previous publications, considering a balanced sample distribution as unproblematic when using ComBat. We do not claim completeness in terms of reporting all technical conditions and possible solutions of the occurring problems as we approach the problem from a clinician’s perspective and not from that of a computer scientist. With our approach of simulating data, we provide readers with a simple method to assess the probability of false positive findings in DNAm microarray data analysis pipelines.