Journal: Computational and Structural Biotechnology Journal

Article Title: i-Modern: Integrated multi-omics network model identifies potential therapeutic targets in glioma by deep learning with interpretability

doi: 10.1016/j.csbj.2022.06.058

Figure Lengend Snippet: Ranking scheme of multi-omics signatures. (A) Copy number variation signatures ranking. (B) Gene expression signatures ranking. (C) DNA methylation signatures ranking. (D) miRNA expression signatures ranking. (E) Protein expression signatures ranking. (F) Somatic mutation signatures ranking.

Article Snippet: We obtained multi-omics glioma datasets, including RNA sequencing data (TPM normalized gene expression quantification), protein expression data (Reverse Phase Protein Array RPPA), miRNA-seq expression data (reads per million for miRNA mapping to miRbase 20), DNA methylation data (Infinium HumanMethylation450 BeadChip), copy number variation data (Affymetrix SNP Array 6.0) and somatic mutation data (DNA sequencing).

Techniques: Biomarker Discovery, Gene Expression, DNA Methylation Assay, Expressing, Mutagenesis

Journal: PLoS ONE

Article Title: DNA Methylation in Multiple Myeloma Is Weakly Associated with Gene Transcription

doi: 10.1371/journal.pone.0052626

Figure Lengend Snippet: (A) Pearson correlation was used to measure linear relationships between DNA methylation and gene expression levels for 1505 CpG probes represented on the GoldenGate Methylation BeadArray. The panels represent examples of a gene with high (left) and low (right) Pearson correlation coefficients when analyzing DNA methylation levels (x axis) against gene expression levels (y axis). (B) A discretization approach was used to classify samples into methylated (M) or unmethylated (U) groups based on the mean ( μ ) methylation value and standard deviation ( σ ) of a given probe. Statistically significant gene expression differences between M and U groups indicated a methylation-expression correlation for the gene in question.

Article Snippet: For these approaches we used DNA methylation data obtained with the GoldenGate BeadArray technology along with corresponding array-based gene expression data from 193 human MM samples.

Techniques: DNA Methylation Assay, Expressing, Methylation, Standard Deviation

Journal: PLoS ONE

Article Title: DNA Methylation in Multiple Myeloma Is Weakly Associated with Gene Transcription

doi: 10.1371/journal.pone.0052626

Figure Lengend Snippet: Box plots represent gene expression levels generated by either microarray or qRT-PCR. Data are shown for samples classified as U or M based on the methylation status of p16 (A), DLC1 (B), IGF1R (C), or IL17RB (D). For microarray data, probe intensities are plotted on the y-axis. Relative fold-change differences are plotted for data generated by qRT-PCR. The number of samples in each group is displayed above each plot. The GoldenGate BeadArray probe names are indicated above each pair of box plots.

Article Snippet: For these approaches we used DNA methylation data obtained with the GoldenGate BeadArray technology along with corresponding array-based gene expression data from 193 human MM samples.

Techniques: Expressing, Generated, Microarray, Quantitative RT-PCR, Methylation

Journal: bioRxiv

Article Title: The DNA methylation landscape of glioblastoma disease progression shows extensive heterogeneity in time and space

doi: 10.1101/173864

Figure Lengend Snippet: Integrative analysis of longitudinal DNA methylation data (RRBS) with matched magnetic resonance (MR) imaging data (morphology, segmentation), clinical annotation data (e.g., treatment, progression, IDH mutation status), and histopathological data (segmentation, morphology, immunohistochemistry) using statistical methods and machine learning. TMZ: Temozolomide; RTX: Radiation therapy; PC: Palliative care. Patient cohort overview summarizing the disease courses of 112 primary glioblastoma patients with IDH- wildtype status, ordered by time of first surgery. DNA methylation profiles for primary and recurring tumors at three relevant gene loci ( BCL2L11 , SFRP2 , and MGMT ). Genes and ENCODE histone H3K27ac tracks were obtained from the UCSC Genome Browser. DNA methylation levels at CpGs indicative of the CpG island methylator phenotype (CIMP), shown separately for IDH mutated control samples (which are CIMP-positive) and the IDH wildtype primary glioblastoma samples from the study cohort (which are CIMP-negative). The fold change of DNA methylation levels between IDH mutated and wildtype samples is indicated based on data from a previous study .

Article Snippet: Classifiers were trained and evaluated on Infinium 27k DNA methylation data for glioblastoma tumors obtained from the TCGA data portal ( https://portal.gdc.cancer.gov/ ).

Techniques: DNA Methylation Assay, Imaging, Mutagenesis, Immunohistochemistry, Control

Journal: bioRxiv

Article Title: The DNA methylation landscape of glioblastoma disease progression shows extensive heterogeneity in time and space

doi: 10.1101/173864

Figure Lengend Snippet: A Overview of the clinical centers that contributed to this study. The numbers of IDH wildtype and IDH mutated patients are indicated for each center. B Scatterplots summarizing the RRBS sequencing data. The proportion of randomly fragmented reads (i.e., reads not starting with the expected RRBS restriction sites) reflect the degree of pre-fragmentation of the input DNA. The different sample types (FFPE: formalin-fixed paraffin-embedded; Frozen: fresh-frozen, RCL: ethanol-based conservation) are indicated by color. C DNA methylation levels of methylated and unmethylated synthetic spike-in control sequences. Dashed lines indicate DNA methylation levels of 5% and 95%. D, E Distribution of DNA methylation levels across different genomic regions (covered by more than 10 reads per CpG) and for the different sample types (D) and quality tiers (E) defined by the number of unique CpGs detected in each RRBS library (tier 1: more than 3 million; tier 2: between 2 and 3 million; tier 3: between 1 and 2 million; tier 4: below 1 million). F Scatterplots depicting the relationship of DNA methylation levels in 5-kilobase tiling regions (containing more than 25 CpGs and covered by more than 10 reads per CpG) between primary and recurring tumors for the three different sample types. r: Pearson correlation. G Mean MGMT promotor methylation levels averaged across two CpGs (cg12434587 and cg12981137) . Error bars indicate the maximum and minimum detected methylation levels in each samples. The dashed line indicates the threshold (36%) below which samples are considered unmethylated. H Kaplan-Meier plots displaying progression-free survival and overall survival probabilities over time for patients stratified by MGMT promoter methylation status as depicted in panel G.

Article Snippet: Classifiers were trained and evaluated on Infinium 27k DNA methylation data for glioblastoma tumors obtained from the TCGA data portal ( https://portal.gdc.cancer.gov/ ).

Techniques: Sequencing, Formalin-fixed Paraffin-Embedded, DNA Methylation Assay, Methylation, Control

Journal: bioRxiv

Article Title: The DNA methylation landscape of glioblastoma disease progression shows extensive heterogeneity in time and space

doi: 10.1101/173864

Figure Lengend Snippet: Overview of the machine learning approach for classifying tumor samples by their transcriptional subtypes using DNA methylation data. Classifiers were trained on DNA methylation data (Infinium 27k assay) of TCGA glioblastoma samples with known transcriptional subtype, using only CpGs shared by RRBS. All classifiers were evaluated by tenfold cross-validation on the TCGA samples and then applied to the RRBS profiles, predicting class probabilities that indicate the relative contribution of each transcriptional subtype. Transcriptional subtype heterogeneity within cohort samples, as indicated by class probabilities of the subtype classifier. Samples are grouped and ordered by their dominant subtype. Distribution of class probabilities across different regions of the same tumor (indicated by Roman numbers) and across different surgeries (indicated by Arabic numbers) for two patients with multisector samples. Riverplot depicting transitions in the predicted transcriptional subtype between primary and recurring tumors. The number of samples in each state is indicated. Only patients whose primary and recurring tumors were classified with high accuracy (ROC AUC > 0.8) were included in this analysis. Kaplan-Meier plots displaying progression-free survival and overall survival probabilities over time for patients stratified by predicted transcriptional subtypes (left) and switching from a non-mesenchymal to mesenchymal subtype during disease progression (right). Only tumor samples that were classified with high accuracy (ROC AUC > 0.8) were included in this analysis. Heatmap displaying the DNA methylation levels of the most differential CpGs between the three transcriptional subtypes. Only tumor samples that were classified with high class probabilities (>0.8) were included in this analysis. LOLA region-set enrichment analysis of differentially methylated CpGs between the different transcriptional subtypes (binned into 1-kilobase tiling regions). Adjusted p-values (Benjamini & Yekutieli method) are displayed for all significantly (adjusted p-value < 0.05) enriched region sets (binding sites, x-axis) measured in astrocytes or embryonic stem cells (ESCs). Schematic depicting the calculation of ‘DNA methylation inferred regulatory activity’ (MIRA) scores. DNA methylation profiles are combined across centered genomic regions of interest (e.g., transcription factor binding sites) for each sample and each region set. The MIRA score is then calculated as the ratio of DNA methylation levels at the flank to DNA methylation levels at the center (binding site) of the combined DNA methylation profile. High MIRA scores therefore reflect local demethylation at the binding site, which indicates high regulatory activity of the respective factor. DNA methylation profiles (upper row) and corresponding MIRA scores (lower row) for three region sets enriched in CpGs that are hypomethylated in the mesenchymal subtype (CTCF binding in astrocytes, EZH2 binding in astrocytes, and KDMA binding in ESCs) as well as three region sets of key regulators of pluripotency measured in ESCs (POUF1, NANOG, SOX2). The significance of differences between the three transcriptional subtypes was assessed using a two-sided Wilcoxon rank sum test: * p-value < 0.05, ** p-value < 0.01, *** p-value < 0.001, ns: not significant.

Article Snippet: Classifiers were trained and evaluated on Infinium 27k DNA methylation data for glioblastoma tumors obtained from the TCGA data portal ( https://portal.gdc.cancer.gov/ ).

Techniques: DNA Methylation Assay, Biomarker Discovery, Methylation, Binding Assay, Activity Assay

Journal: bioRxiv

Article Title: The DNA methylation landscape of glioblastoma disease progression shows extensive heterogeneity in time and space

doi: 10.1101/173864

Figure Lengend Snippet: A Comparison of tumor-infiltrating immune cell levels between different transcriptional subtypes as measured by quantitative immunohistochemistry for the indicated marker proteins. B Immunohistochemical stainings for FOXP3 and CD45ro in selected samples assigned to each of the three transcriptional subtypes. C Kaplan-Meier plots displaying progression-free survival and overall survival probabilities over time for patients stratified according to the level of CD163-positive and CD68-positive immune cell infiltration in their primary and recurring tumors. D, E Differences in the relative proportion of tumor-infiltrating pro-inflammatory (D) and anti-inflammatory or neutral (E) immune cells between tumor samples originating from the patients’ first surgery (primary tumor), second surgery (recurring tumor), or third surgery. F Comparative immunohistochemical stainings between primary and recurring tumors for three selected markers (CD68, CD8, CD163). G Comparison of the levels of tumor-infiltrating immune cells (cells positive for CD3, CD8, or CD68) and proliferating cells (MIB-positive cells) between the different progression types based on magnetic resonance (MR) imaging: Classic T1 (claT1), cT1 relapse / flare-up (cT1), and T2 diffuse (T2). Primary (left panel) and recurring (right panel) tumors were analyzed separately. H ROC curves for the DNA methylation based prediction of immune cell infiltration levels, as determined by leave-one-out cross-validation. ROC curves and the ROC area under curve (AUC) are indicated for the actual prediction (blue) and for background predictions with randomly shuffled labels (grey). All significance tests comparing groups of samples in this figure were performed using a two-sided Wilcoxon rank sum test: * p-value < 0.05, ** p-value < 0.01, *** p-value < 0.001, ns: not significant.

Article Snippet: Classifiers were trained and evaluated on Infinium 27k DNA methylation data for glioblastoma tumors obtained from the TCGA data portal ( https://portal.gdc.cancer.gov/ ).

Techniques: Comparison, Immunohistochemistry, Marker, Immunohistochemical staining, Imaging, DNA Methylation Assay, Biomarker Discovery

Journal: bioRxiv

Article Title: The DNA methylation landscape of glioblastoma disease progression shows extensive heterogeneity in time and space

doi: 10.1101/173864

Figure Lengend Snippet: T1-contrast enhanced and T2/FLAIR MR sequences at each follow-up visit illustrating the three MR imaging progression types in this cohort (cT1 relapse / flare-up, classic T1, T2 diffuse). ‘ * ’: tumor recurrence. Kaplan-Meier plots displaying progression-free survival and overall survival probabilities over time for patients stratified according to their MR imaging progression types. Schematic illustrating the machine learning approach (including the data pre-processing) used to assess the predictability of various tumor properties from RRBS DNA methylation data. White squares indicate missing values; grey squares indicate imputed values. ROC curves showing high prediction accuracy based on DNA methylation data for two features with high expected predictability (IDH mutation status, patient sex). ROC curves evaluating the DNA methylation based prediction of several histopathologic tumor properties. Hierarchical clustering based on the column-scaled DNA methylation values of the most predictive features (5-kilobase tiling regions) as identified by the machine leaning classifiers built to predict the infiltration levels of the indicated immune cell types (cells positive for CD163, CD68, CD45ro, CD3, or CD8) or the extent of indicated tumor properties (cells positive for CD34 or HLA-DR).

Article Snippet: Classifiers were trained and evaluated on Infinium 27k DNA methylation data for glioblastoma tumors obtained from the TCGA data portal ( https://portal.gdc.cancer.gov/ ).

Techniques: Imaging, DNA Methylation Assay, Mutagenesis

Journal: bioRxiv

Article Title: The DNA methylation landscape of glioblastoma disease progression shows extensive heterogeneity in time and space

doi: 10.1101/173864

Figure Lengend Snippet: Comparison of the fraction of proliferating (MIB-positive) cells between first surgery (primary tumor), second surgery (recurring tumor), and third surgery. Kaplan-Meier plots displaying progression-free survival and overall survival probabilities over time for patients stratified according to the fraction of proliferating (MIB-positive) cells in the primary and recurring tumor. ROC curves for the DNA methylation based prediction of levels of proliferating (MIB positive) cells, as determined by leave-one-out cross-validation. Hierarchical clustering based on column-scaled DNA methylation levels of the most predictive genomic regions from the classifier predicting the fraction of proliferating (MIB-positive) cells (5-kilobase tiling regions). Distribution of DNA methylation levels of the most predictive genomic regions from the classifier predicting the fraction of proliferating (MIB-positive) cells, displayed separately for samples with high (red) or low (blue) fractions and for positively (top) or negatively (bottom) associated features as identified by the classifier. Comparison of the average nuclear eccentricity (AVG) and its coefficient of variation (COV) between tumors that shift to a sarcoma-like phenotype during disease progression and those that retain a stable histological phenotype. Hematoxylin and eosin stains of matched primary and recurring tumors, illustrating the morphological changes observed when tumors shift to a sarcoma-like phenotype during disease progression. Comparison of additional tumor properties between tumors that shift to a sarcoma-like phenotype during disease progression and those that retain a stable histological phenotype. ROC curves for the DNA methylation based prediction of average nuclear eccentricity and its coefficient of variation. Kaplan-Meier plots displaying progression-free survival and overall survival probabilities over time for patients stratified according to whether their tumors shift to a sarcoma-like phenotype during disease progression (green) or retain a stable histological phenotype (orange). All significance tests comparing groups of samples in this figure were performed using a two-sided Wilcoxon rank sum test: * p-value < 0.05, ** p-value < 0.01, *** p-value < 0.001, ns: not significant.

Article Snippet: Classifiers were trained and evaluated on Infinium 27k DNA methylation data for glioblastoma tumors obtained from the TCGA data portal ( https://portal.gdc.cancer.gov/ ).

Techniques: Comparison, DNA Methylation Assay, Biomarker Discovery

Journal: bioRxiv

Article Title: The DNA methylation landscape of glioblastoma disease progression shows extensive heterogeneity in time and space

doi: 10.1101/173864

Figure Lengend Snippet: Scatterplots displaying the relationship between PCR enrichment cycles during RRBS library preparation and the indicated measures of DNA methylation heterogeneity. In order to reduce the effect of technical variability, we limited the analysis of epigenomic heterogeneity to samples that fall into a defined narrow range of PCR enrichment cycles (13-15 cycles, indicated by black boxes). Degree of epi-allelic shifting between normal brain control and primary or recurring tumors, as well as between primary and recurring tumors measured by the relative number of loci that show high changes in epi-allele composition (EPM: eloci per million assessed loci). *** : p-value < 0.001 (two-sided Wilcoxon rank sum test) Kaplan-Meier plots displaying progression-free survival and overall survival probabilities over time for patients stratified according to their degree of epi-allelic shifting (as measured by EPM) between primary and recurring tumors. Correlation between epi-allelic shifting during progression (as measured by EPM) and the time between first and second surgery. r: Pearson correlation.

Article Snippet: Classifiers were trained and evaluated on Infinium 27k DNA methylation data for glioblastoma tumors obtained from the TCGA data portal ( https://portal.gdc.cancer.gov/ ).

Techniques: DNA Methylation Assay, Control

Journal: bioRxiv

Article Title: The DNA methylation landscape of glioblastoma disease progression shows extensive heterogeneity in time and space

doi: 10.1101/173864

Figure Lengend Snippet: Scatterplot depicting the relationship of promoter DNA methylation between primary and recurring tumors. Promoters that were differentially methylated between primary and recurring tumors in at least 5 patients are highlighted (DNA methylation difference greater than 75%, adjusted p-value below 0.001, and average RRBS read coverage greater than 20 reads). r: Pearson correlation. Barplots (top) depicting the number of patients that show significant gain or loss of DNA methylation in the differentially methylated promoters highlighted in panel A; scatterplots and line plots (bottom) showing the change in DNA methylation associated with disease progression (measured as percentage points, pp) for patients following the cohort-trend (red) or not (blue). Trend lines were calculated using the loess method. Definition of “trend” and “anti-trend” patients based on the Manhattan distance between the maximal trend at differentially methylated promoters (DNA methylation values of 0% or 100%) and the observed difference in DNA methylation for each patient. “Trend” patients are those whose DNA methylation profiles are similar to the maximal trend (low normalized Manhattan distance); “Anti-trend” patients are those whose methylation profiles are most different from the maximal trend (high normalized Manhattan distance). Pathway enrichment analysis of those genes that recurrently lose DNA methylation during disease progression and those that recurrently gain DNA methylation during disease progression. Kaplan-Meier plots displaying progression-free survival and overall survival probabilities over time for patients stratified by whether they followed the cohort trend of differential promoter DNA methylation (trend patients) or not (anti-trend patients), according to the definition in panel C. Kaplan-Meier plots displaying progression-free survival and overall survival probabilities over time for patients stratified into the top-30% patients with increasing or decreasing average DNA methylation levels at the promoters of Wnt signaling genes during disease progression.

Article Snippet: Classifiers were trained and evaluated on Infinium 27k DNA methylation data for glioblastoma tumors obtained from the TCGA data portal ( https://portal.gdc.cancer.gov/ ).

Techniques: DNA Methylation Assay, Methylation, Biomarker Discovery

Journal: Aging Cell

Article Title: Distinct chromatin signatures of DNA hypomethylation in aging and cancer

doi: 10.1111/acel.12744

Figure Lengend Snippet: DNA methylation changes in aging and cancer. (a) Stacked barplots indicating total number of dmCpGs detected in cancer, aging, and aging cancer tissues. (b) Hierarchical clustering and heatmaps including the 1,000 most significant dmCpGs for breast cancer and aging analyses. Beta‐values of DNA methylation are displayed from zero (green) to one (red). (c) Boxplots comparing the magnitude of M‐values of methylation changes in cancer and aging. All differences are statistically significant (Wilcoxon tests, all p < .05, Table ). (d) Scatterplots indicating a correlation of chronological age with Horvath's predicted age in normal and cancer samples. Pearson's product‐moment correlation coefficient (cor) is indicated, and linear fit lines are added to help with data interpretation

Article Snippet: To address this issue, we analyzed TCGA DNA methylation data from a total of 2,311 samples, including control and cancer cases from patients with breast, kidney, thyroid, skin, brain, and lung tumors and healthy blood, and integrated the results with histone, chromatin state, and transcription factor binding site data from the NIH Roadmap Epigenomics and ENCODE projects.

Techniques: DNA Methylation Assay, Methylation

Journal: Aging Cell

Article Title: Distinct chromatin signatures of DNA hypomethylation in aging and cancer

doi: 10.1111/acel.12744

Figure Lengend Snippet: DNA methylation signatures of aging and cancer. (a) Stacked barplots indicating, in blue, the total number of hyper‐ and hypomethylated dmCpGs detected between all of the tissues analyzed in cancer and aging. Of these, the proportion of dmCpGs not shared between any tissues (unique, orange) or those shared by two or more tissues (shared, green) is shown in the adjacent barplot. (b) Venn diagrams depicting the number of differentially hyper‐ and hypomethylated CpGs in aging and cancer shared by the different tissues. (c) Venn diagrams showing the number and overlap of total nonredundant hyper‐ and hypomethylated dmCpGs detected in cancer and aging. (d) Heatmaps showing pairwise comparisons between sets of probes: in green, Jaccard Indices; in red, odd ratios (all enrichment Fisher's tests p < .001)

Article Snippet: To address this issue, we analyzed TCGA DNA methylation data from a total of 2,311 samples, including control and cancer cases from patients with breast, kidney, thyroid, skin, brain, and lung tumors and healthy blood, and integrated the results with histone, chromatin state, and transcription factor binding site data from the NIH Roadmap Epigenomics and ENCODE projects.

Techniques: DNA Methylation Assay

Journal: Aging Cell

Article Title: Distinct chromatin signatures of DNA hypomethylation in aging and cancer

doi: 10.1111/acel.12744

Figure Lengend Snippet: Relationships between DNA methylation and gene expression in aging and cancer. (a) Venn diagrams illustrating the overlap between DEGs in aging and cancer in the KIRC dataset (see Table for DEG lists). (b) Boxplot depicting the DNA methylation β‐values of the CpG cg19442915 in old and young individuals ( n = 5) from the KIRC aging condition. (c) Boxplot showing the gene expression values (RSEM) of the CKM gene in old and young individuals ( n = 5) from the aging condition of the KIRC dataset. (d) Scatterplot showing the Spearman correlation between DNA methylation (cg19442915) and gene expression (CKM gene) in 18 normal kidney samples. Colored dots indicate old or young individuals used for the aforementioned aging comparisons. (e) Histograms representing the number of pairwise correlations that are contained in a given correlation window (from 1 to −1) obtained as the result of computing the correlation between β‐values of dmCpGs identified in cancer or aging and gene expression levels (RSEM) of genes expressed in the KIRC dataset. The number of dmCpGs used for each of the comparisons is indicated at the bottom. (f) Barplots depicting the number of total (blue) and unique (orange) dmCpG‐gene pairs identified in the previous analysis which displayed correlation scores above 0.9 (pos) or below −0.9 (neg) in cancer (top) or in aging (bottom) conditions. (g) Stacked barplots indicating relative distribution of unique dmCpGs obtained from the previous correlations according to their CpG island status. (h) Stacked barplots indicating relative distribution of unique dmCpGs obtained from the previous correlations according to their gene location status. (i) Venn diagrams illustrating the overlap between dmCpGs identified in aging and in cancer which displayed strong positive or negative correlations (>0.9 or <−0.9) with genes expressed in the normal kidney dataset

Article Snippet: To address this issue, we analyzed TCGA DNA methylation data from a total of 2,311 samples, including control and cancer cases from patients with breast, kidney, thyroid, skin, brain, and lung tumors and healthy blood, and integrated the results with histone, chromatin state, and transcription factor binding site data from the NIH Roadmap Epigenomics and ENCODE projects.

Techniques: DNA Methylation Assay, Gene Expression

Journal: Aging Cell

Article Title: Distinct chromatin signatures of DNA hypomethylation in aging and cancer

doi: 10.1111/acel.12744

Figure Lengend Snippet: Description of sample groups and dmCpGs obtained in the analyses

Article Snippet: To address this issue, we analyzed TCGA DNA methylation data from a total of 2,311 samples, including control and cancer cases from patients with breast, kidney, thyroid, skin, brain, and lung tumors and healthy blood, and integrated the results with histone, chromatin state, and transcription factor binding site data from the NIH Roadmap Epigenomics and ENCODE projects.

Techniques:

Journal: Molecular Oncology

Article Title: Allele‐specific proximal promoter hypomethylation of the telomerase reverse transcriptase gene ( TERT ) associates with TERT expression in multiple cancers

doi: 10.1002/1878-0261.12786

Figure Lengend Snippet: Cells with monoallelic expression of TERT have allelic methylation of the proximal TERT promoter. (A) Extent of allelic methylation across the TERT gene, which is transcribed from right to left. Relative allelic methylation was measured by calculating the difference between the mean and mode values of raw read Bis‐Seq CpG methylation data, where greater levels suggest greater allelic methylation behavior (see Fig. for examples). Positions included contained 3–6 CpGs per read and coverage of ≥ 5 reads per cell line. Error bars represent standard error of the mean. n represents the number of cancer cell lines. * P ≤ 0.05, where statistical analysis was performed using 2‐tailed Student's t ‐test with unequal variance. (B) Bisulfite conversion cloning data from genomic DNA of select CpGs flanking the TERT transcription start site (5:1295138–1295413, spanning 33 CpGs). Each row represents a different clone (or genome copy, or allele) and each circle represents a CpG, with black circles representing a methylated CpG and white circles representing an unmethylated CpG. For chromosomal positions of noted TERT features, see Table .

Article Snippet: Cancer cell line DNA CpG methylation data were downloaded from the Broad Institute's Cancer Cell Line Encyclopedia (CCLE) [ ] ( www.broadinstitute.org/ccle , June 14, 2018, release, TERT gene) for the TERT gene and 1 kb upstream of the translation start site.

Techniques: Expressing, Methylation, CpG Methylation Assay, Cloning

Journal: Molecular Oncology

Article Title: Allele‐specific proximal promoter hypomethylation of the telomerase reverse transcriptase gene ( TERT ) associates with TERT expression in multiple cancers

doi: 10.1002/1878-0261.12786

Figure Lengend Snippet: Decreased TERT promoter methylation associates with histone marks of active transcription and an active exonic SNP. (A) ChIP‐Bis‐Seq of the TERT promoter using an H3ac antibody shows enrichment of unmethylated DNA in the pulled‐down samples (black) relative to the input (gray) in LN‐18 cells. The absence of any bars indicates zero percent methylation. Inclusion criteria for read positions were a greater number of reads in the pull‐down relative to the input and ≥ 10 reads in the pull‐down (mean input coverage was 9 reads; mean pull‐down coverage was 13 reads; P = 0.01 for pull‐down efficiency). (B) Confirmation of long‐range bisulfite conversion PCR enriching for unmethylated or methylated CpGs at the TERT proximal promoter (16 CpGs spanning 5:1295265–1295396; region overlaps with some of the CpGs analyzed in 3A) using unmethylated (gray)‐ or methylated (black)‐specific bisulfite conversion PCR, respectively. PCR products generated a 1448‐bp product including the proximal promoter and the exon 2 SNP analyzed in Panel C. * P ≤ 0.05 (C) Long‐range bisulfite conversion PCR (same PCRs as shown in Panel B) showing representative Sanger sequencing results (upward arrow indicates position of the exon 2 SNP) and graphs of the sequencing results ( n = 2–3 sequenced reactions). ‘Active SNP’ means that the nucleotide at the position of the SNP is the one found in the TERT mRNA transcribed in that cell line. The active SNP was either previously identified in all cell lines or was identified here (Fig. S6). Error bars represent standard error of the mean. * P ≤ 0.01, where statistical analysis was performed using 2‐tailed Student's t ‐test with unequal variance.

Article Snippet: Cancer cell line DNA CpG methylation data were downloaded from the Broad Institute's Cancer Cell Line Encyclopedia (CCLE) [ ] ( www.broadinstitute.org/ccle , June 14, 2018, release, TERT gene) for the TERT gene and 1 kb upstream of the translation start site.

Techniques: Methylation, Generated, Sequencing

Journal: Molecular Oncology

Article Title: Allele‐specific proximal promoter hypomethylation of the telomerase reverse transcriptase gene ( TERT ) associates with TERT expression in multiple cancers

doi: 10.1002/1878-0261.12786

Figure Lengend Snippet: TERT promoter is characterized by conserved upstream hypermethylation and proximal hypomethylation across different cancer tissue types. (A) Bisulfite conversion sequencing (Bis‐Seq) DNA CpG methylation data for 95 positions across the TERT promoter for 23 different cancerous tissues, showing mean values from 833 cancer cell lines ( n represents the number of cell lines per tissue). Colored circles indicate individual CpG sites with statistically significant ( P ≤ 0.005) differences between the tissue and all other tissues. Each graph groups tissues by total mean percent methylated, from most to least methylated (top to bottom, respectively). Each chromosomal position includes data from at least two cell lines for all 23 tissues. (B) Bis‐Seq DNA CpG methylation data for 129 positions across the TERT promoter for 109 cell lines with known allelic expression and activating mutation classifications. Lines had been classified as having wild‐type (WT) monoallelic expression (MAE) of TERT (‘WT MAE’), −124 or −146 C>T activating promoter mutations (‘mutant’), biallelic expression (BAE) of TERT (‘WT BAE’), or alternative lengthening of telomeres (ALT). Each row represents a different cell line. Colors range from red to blue for more to less methylated CpGs, respectively. White represents unavailable data. Each chromosomal position includes data from at least 10 cell lines. (C) Bis‐Seq DNA CpG methylation data for 122 positions across the TERT promoter for 107 cell lines shown in 1B ( n represents the number of cell lines per tissue). Colored circles indicate statistically significant ( P ≤ 0.05) differences in the listed pairwise comparisons, where statistical analysis was performed using 2‐tailed Student's t ‐test with unequal variance. Each chromosomal position includes data from at least two cell lines for all three cell types. See Table for chromosomal positions and Table for cell line data.

Article Snippet: Cancer cell line DNA CpG methylation data were downloaded from the Broad Institute's Cancer Cell Line Encyclopedia (CCLE) [ ] ( www.broadinstitute.org/ccle , June 14, 2018, release, TERT gene) for the TERT gene and 1 kb upstream of the translation start site.

Techniques: Sequencing, CpG Methylation Assay, Methylation, Expressing, Mutagenesis