Journal: bioRxiv

Article Title: Detection of glioma and prognostic subtypes by non-invasive circulating cell-free DNA methylation markers

doi: 10.1101/601245

Figure Lengend Snippet: A) Kaplan-Meier survival curves showing samples separated by IDH status (left) and IDH status combined with 1p/19q co-deletion (right). Tick represents censorship. B) Total serum cfDNA concentration normalized to the genomic size (Genomic Equivalents/ml) in non-tumor, IDH mut and IDH wt samples. C) Heatmap of DNA methylation probes mapped the promoter region of MGMT gene. DNA methylation beta-values are represented as a color gradient from low (blue) to high (red) in C, D, and E. D) Heatmap of DNA methylation of probes that define epigenetically regulated genes in glioma subtypes. Rows represent EReg probes described by Ceccarelli et al., 2016. E) Predictive biomarkers for glioma progression. Rows represent probes that stratify gliomas into risk for aggressive recurrence (deSouza et al., 2018). Each marker was coded as white if methylated and black if unmethylated according to the published cutoffs. F) Stemness index defined by DNA methylation as described by Malta et al. 2018. Samples are stratified by genomic group and by sample type.

Article Snippet: The extracted DNA (30-300 ng) was bisulfite-converted (Zymo EZ DNA methylation Kit; Zymo Research) and profiled using an Illumina Human EPIC array (HM850K), at the USC Epigenome Center, Keck School of Medicine, University of Southern California, Los Angeles, California.

Techniques: Concentration Assay, DNA Methylation Assay, Marker, Methylation

Journal: bioRxiv

Article Title: Detection of glioma and prognostic subtypes by non-invasive circulating cell-free DNA methylation markers

doi: 10.1101/601245

Figure Lengend Snippet: A) Total serum cfDNA concentration normalized to the genomic size (Genomic Equivalents/ml) is represented in two boxplots (non-tumor vs glioma). B) Epigenome-wide cell-free DNA methylation (serum-based) derived from Glioma, Pituitary tumor, Colorectal carcinoma (CRC), and non-tumor patients presented by a similarity method: Principal Components Analysis (PCA). Each dot represents a sample cfDNA methylation (epigenome-wide) and colored based on tissue/tumor of origin. Total percent variance is indicated along all three axes. C) Cell-type CpG methylation-based deconvolution of our patients’ cfDNA methylome is divided into three relevant categories: Neural, Immune and Other cell types. Y-axis represents normalized percent of each cell-type and separately by serum and tissue present in the cfDNA methylome. Eleven sets of boxplots are each divided into two categories; non-tumor (grey) vs glioma (blue). The plots are further divided by tissue (non-tumor-light grey or glioma-light blue) and cfDNA serum (non-tumor-dark grey or glioma-dark blue). D) Published tissue-derived epigenetic signatures from Ceccarelli et al. 2016 (top heatmaps) and Sturm et al. 2012 (bottom heatmaps) are undetectable in Glioma cfDNA methylome. Levels of CpG methylation in our HBTC cohort divided by glioma tissue (left two heatmaps, respectively) vs matching serum cfDNA (right two heatmaps). Columns indicate patients with clinical and molecular annotation tracks listed on top of heatmaps and rows indicates CpG probe. Beta-value (DNA methylation levels) are indicated in the legend from 0 (low CpG methylation) to 1 (high CpG methylation). E) Glioma-tissue epigenome-wide PCA highlights the difference between glioma and non-tumor and between IDH mut and IDH wt patients. TCGA Pan-Glioma DNA methylation data from Ceccarelli et al. 2016 is represented in a Principal Component Analysis to evaluate similarities genome-wide. The first two principal components are plotted using all available DNA methylation data points (~400,000 CpGs, unfiltered). The variance total percentage is labeled along both axes. Fifty-four total non-glioma tissue (grey) were also included in this analysis to highlight the epigenome-wide difference between glioma and non-tumor. The glioma cohort is further divided by available known IDH status and by our recent epigenomic subtypes.

Article Snippet: The extracted DNA (30-300 ng) was bisulfite-converted (Zymo EZ DNA methylation Kit; Zymo Research) and profiled using an Illumina Human EPIC array (HM850K), at the USC Epigenome Center, Keck School of Medicine, University of Southern California, Los Angeles, California.

Techniques: Concentration Assay, DNA Methylation Assay, Derivative Assay, Methylation, CpG Methylation Assay, Genome Wide, Labeling

Journal: bioRxiv

Article Title: Detection of glioma and prognostic subtypes by non-invasive circulating cell-free DNA methylation markers

doi: 10.1101/601245

Figure Lengend Snippet: A) Heatmap of DNA methylation of Glioma-eLB probes (N = 1075 CpG sites). B) DNA methylation levels of Glioma-eLB and previously published probes (Ceccarelli et al., 2016) in the serum and in the tissue of a representative patient. C) Dendrograms of non-tumor cell types and serum using Glioma non-tissue-specific CpGs (N = 186, left) in comparison to tissue-specific (N=384, right) CpGs. D) Glioma-specific tissue-matching eLB (N=384) was used to subset the published primary tumor tissue DNA methylation (N=33 tumor types) and using t-SNE (t-distributed stochastic neighbour embedding) dimensionality reduction to visualize the similarities of each sample. As expected, each primary tumor type (circles) clusters with its known cell-of-origin. Serum cfDNA methylation of our patients cohort (triangles) clusters with the primary glioma tissue DNA methylation profiles.

Article Snippet: The extracted DNA (30-300 ng) was bisulfite-converted (Zymo EZ DNA methylation Kit; Zymo Research) and profiled using an Illumina Human EPIC array (HM850K), at the USC Epigenome Center, Keck School of Medicine, University of Southern California, Los Angeles, California.

Techniques: DNA Methylation Assay, Methylation

Journal: bioRxiv

Article Title: Detection of glioma and prognostic subtypes by non-invasive circulating cell-free DNA methylation markers

doi: 10.1101/601245

Figure Lengend Snippet: A) Epigenome-wide mean DNA methylation across our patient cohort’s serum cfDNA methylation (y-axis: three nontumor samples and x-axis: 22 glioma-patient-derived serum samples). Non-significant CpG methylation probes (p-value > 5%) are condensed into a density heatmap by calculating the 2D kernel density estimation to the power of 0.1. Identified cfDNA methylation signatures associated with glioma patients (N = 1,075) are selected by different p-values (black, p-value < 0.05; blue, and red, p-value < 0.001). Glioma-eLB signatures (p-value < 0.001) are further divided into CpGs that are measurable (Glioma tissue specific, red) or not measurable in the matching glioma tissue (Non-tissue specific, blue). B) Principal Component Analysis using the Glioma-eLB signatures as input. Serum methylome from glioma, pituitary tumors, CRC and non-tumor samples are represented. C) Glioma-specific tissue-matching eLB (N=384) was used to subset the published primary tumor tissue DNA methylation and using t-SNE (t-distributed stochastic neighbour embedding) dimensionality reduction to visualize the similarities of each sample. As expected, each primary tumor type (circles) clusters with its known cell-of-origin. Serum cfDNA methylation of our patients cohort (triangles) clusters with the primary glioma tissue DNA methylation profiles. D) Dendrogram of non-tumor cell types in comparison to Glioma-eLB (N=384). Based on the mean DNA methylation across each cell type, this dendrogram shows that glioma serum cfDNA clusters with relevant immune cell-types along with glial-derived cells and bulk brain (non-tumor) samples. E) Machine learning (ML) application (Random Forest) using our defined Glioma tissue-specific eLB to classify tumors and available cfDNA methylation (serum or plasma) derived from tumor patients, patients with metastasis of unknown primary, non-tumor conditions (e.g., sepsis, pancreatic islet transplantation recipient) and non-tumor/non-diseased cell-free DNA. Y-axis represents the ML similarity index based on Glioma-eLB signatures averaged across 1000 iterations. Zero indicates low probability of a glioma, while 1 indicates high probability of a sample being a glioma. Dash line indicates cutoff to determine glioma classification. F) Receiver operating characteristic curve derived from the average across 1000 cross validation. Specificity and sensitivity calculated at 0.05 increments (N=21) from 0 to 1 Glioma-eLB signature index.

Article Snippet: The extracted DNA (30-300 ng) was bisulfite-converted (Zymo EZ DNA methylation Kit; Zymo Research) and profiled using an Illumina Human EPIC array (HM850K), at the USC Epigenome Center, Keck School of Medicine, University of Southern California, Los Angeles, California.

Techniques: DNA Methylation Assay, Methylation, Derivative Assay, CpG Methylation Assay, Transplantation Assay

Journal: bioRxiv

Article Title: Detection of glioma and prognostic subtypes by non-invasive circulating cell-free DNA methylation markers

doi: 10.1101/601245

Figure Lengend Snippet: A) Mean DNA methylation of 750,000 CpG across 15 IDH mut-patient derived serum (x-axis) vs 7 IDH wt-patient derived serum (y-axis). Non-significant CpG methylation probes are condensed into a density heatmap by calculating the 2D kernel density estimation to the power of 0.1. Proposed prognostic glioma-specific eLB (N=1,075) selected by p-values are represented by colored dots (black, p-value < 0.05; blue, red and yellow, p-value < 0.01). Similarities across selected tumor tissue and serum, using IDH mut-tissue specific eLB levels (red circles) and IDH wt- tissue specific eLB levels (yellow circles). B-C) Similarities across Pan-Glioma tissue (N = 259 IDH mut and 160 IDH wt) and IDH glioma cfDNA methylation (N = 15 IDH mut, 7 IDH wt), using IDH mut-tissue specific eLB signatures B) and IDH wt-tissue specific eLB signatures C) as input in a t-SNE analysis to visualize the similarities by sample. Circles represent primary tissue and triangles represents tumor serum cfDNA. Red indicates IDH mut and purple indicates IDH wt. D-E) scatter plot between DNA methylation (x-axis) and Gene expression (y-axis) for all pan-glioma primary tumor tissue. 2D kernel density indicates all glioma samples divided by IDH status (purple = IDH wt, orange = IDH mut). Circles indicate the HBTC primary tumor tissue DNA methylation and expression values. D) DNA methylation and expression scatter plot for promoter CpG associated with CXCR6 . E) DNA methylation and expression scatter plot for promoter CpG associated with PVT1 .

Article Snippet: The extracted DNA (30-300 ng) was bisulfite-converted (Zymo EZ DNA methylation Kit; Zymo Research) and profiled using an Illumina Human EPIC array (HM850K), at the USC Epigenome Center, Keck School of Medicine, University of Southern California, Los Angeles, California.

Techniques: DNA Methylation Assay, Derivative Assay, CpG Methylation Assay, Methylation, Expressing

Journal: bioRxiv

Article Title: Detection of glioma and prognostic subtypes by non-invasive circulating cell-free DNA methylation markers

doi: 10.1101/601245

Figure Lengend Snippet: A) Heatmap of DNA methylation of IDH-eLB probes (N = 114 IDH mut-eLB and 124 IDH wt-eLB CpG sites). B) Comparison between DNA methylation levels of glioma tissue (y-axis) and serum (x-axis) of one representative IDH mut glioma patient on the left and one representative IDH wt glioma patient on the right. C-D) IDH -eLB (N=114 IDH mut-eLB and 124 IDH wt-eLB CpG sites) was used to subset the published primary tumor tissue DNA methylation data and using t-SNE (t-distributed stochastic neighbour embedding) dimensionality reduction to visualize the similarities of each sample. C) t-SNE using IDH mut-eLB CpGs on the left and IDH wt-eLB CpGs on right with primary TCGA tumor tissue from tumor types (N=9) with known IDH mutation. As expected, each primary tumor type (circles) clusters with its known cell-of-origin. Serum cfDNA methylation of our IDH mut patients cohort (triangles) clusters with the IDH mut primary glioma tissue and serum cfDNA methylation of our IDH wt patients cohort (triangles) clusters with the IDH wt primary glioma tissue. D) t-SNE using IDH mut-eLB CpGs on the left and IDH wt-eLB CpGs on right with primary TCGA tumor tissue (N=33 tumor types). E) Odds-ratio for the frequencies of IDH mut-eLB probes (left) and IDH wt-eLB (right), respectively, that overlap a particular molecular feature relative to the expected genome-wide distribution of the methylation platform.

Article Snippet: The extracted DNA (30-300 ng) was bisulfite-converted (Zymo EZ DNA methylation Kit; Zymo Research) and profiled using an Illumina Human EPIC array (HM850K), at the USC Epigenome Center, Keck School of Medicine, University of Southern California, Los Angeles, California.

Techniques: DNA Methylation Assay, Mutagenesis, Methylation, Genome Wide

Journal: bioRxiv

Article Title: Detection of glioma and prognostic subtypes by non-invasive circulating cell-free DNA methylation markers

doi: 10.1101/601245

Figure Lengend Snippet: A) DNA methylation (x-axis) and expression (y-axis) scatter plot for promoter CpG associated with the corresponding gene. Each dot represents a sample. Red represents IDH mut glioma tissue samples, dark purple represents IDH wt glioma tissue samples, orange represents IDH mut glioma serum cfDNA and light purple IDH wt glioma serum cfDNA.

Article Snippet: The extracted DNA (30-300 ng) was bisulfite-converted (Zymo EZ DNA methylation Kit; Zymo Research) and profiled using an Illumina Human EPIC array (HM850K), at the USC Epigenome Center, Keck School of Medicine, University of Southern California, Los Angeles, California.

Techniques: DNA Methylation Assay, Expressing

Journal: bioRxiv

Article Title: Detection of glioma and prognostic subtypes by non-invasive circulating cell-free DNA methylation markers

doi: 10.1101/601245

Figure Lengend Snippet: Real time diagnosis and, surveillance of early tumor progression or recurrence. A) Steps used to generate the eLB and the application of a machine learning (ML) model to predict glioma and glioma subtypes. B) At specific intervals, currently determined by magnetic resonance imaging (MRI) visits, patients’ whole-blood is collected and serum/plasma is immediately processed. cfDNA is isolated and profiled using DNA methylation microarray (profile >700,000 CpGs across the entire human genome) and entered into a ML algorithm to generate a index (Glioma-index). C) According to the predetermined index threshold (0.6, sensitivity: 98%; specificity: 99%), the new sample is classified as glioma or non-glioma and according to IDH status (mutant or wildtype). The discovered glioma specific eLB (N = 1,075) can be used to complement current clinical diagnostic and monitoring events. Prognostic IDH -eLB could be used at time of diagnosis and for monitoring during active treatment through survivorship care. Complementing the MRI findings, eLB could improve detection, reduce false-positive and increase tumor identification (glioma vs necrosis vs non glioma conditions), assess treatment outcome and help tailor treatment options for patients with glioma. eLB could also foster early detection of glioma progression to improve treatment outcomes. Future clinical trials are needed to evaluate the robustness of our eLB signatures for clinical application.

Article Snippet: The extracted DNA (30-300 ng) was bisulfite-converted (Zymo EZ DNA methylation Kit; Zymo Research) and profiled using an Illumina Human EPIC array (HM850K), at the USC Epigenome Center, Keck School of Medicine, University of Southern California, Los Angeles, California.

Techniques: Magnetic Resonance Imaging, Isolation, DNA Methylation Assay, Microarray, Mutagenesis, Diagnostic Assay

Journal: Environment International

Article Title: Waterpipe and cigarette epigenome analysis reveals markers implicated in addiction and smoking type inference

doi: 10.1016/j.envint.2023.108260

Figure Lengend Snippet: Characteristics of never, cigarette and waterpipe smokers in the discovery sets analyzed by microarray (A-B) and validation set analyzed by pyrosequencing (C).

Article Snippet: The Discovery set constituted of DNA methylome-wide array profiling by Infinium Methylation EPIC microarray (850 K array) of 72 blood samples taken from each of the never, current cigarette-only and current waterpipe-only smokers ( ; A ).

Techniques: Microarray, Activity Assay

Journal: Scientific Reports

Article Title: Epigenome-wide association study of diabetic chronic kidney disease progression in the Korean population: the KNOW-CKD study

doi: 10.1038/s41598-023-35485-x

Figure Lengend Snippet: DNA methylation analysis by epigenome-wide association study as discovery and pyrosequencing as validation for the progression of chronic kidney disease (CKD) in diabetic CKD patients. Beeswarm and box plots shows the DNA methylation values of two CpG sites. ( A ) The M-values and beta-values of epigenome-wide association study based on EPIC BeadChip for (a), cg02990553 on AGTR1 (b), and cg10297223 on KRT28 . ( B ) The percentage of differentially methylated CpG sites using pyrosequencing were generated for (a), cg02990553 on AGTR1 (b), and cg10297223 on KRT28 .

Article Snippet: The annotation was performed by an Illumina Infinium MethylationEPIC BeadChip (EPIC chip), which is a microarray platform designed to DNA methylation across over 860,000 CpG sites in human genome.

Techniques: DNA Methylation Assay, Methylation, Generated

Journal: Frontiers in Public Health

Article Title: Analysis of Epigenetic Age Predictors in Pain-Related Conditions

doi: 10.3389/fpubh.2020.00172

Figure Lengend Snippet: List of variables calculated by the new DNA methylation age calculator available online at https://dnamage.genetics.ucla.edu/ .

Article Snippet: Raw data files ( .idat format) from the three studies were downloaded and separately pre-processed using minfi package within Rstudio software (version 3.5.1) in Linux environment. minfi package provides tools for the analysis of Infinium DNA Methylation microarrays and can handle both 450k and EPIC arrays ( , ).

Techniques: DNA Methylation Assay, Methylation