Differential Methylation

Introduction: Differential Methylation of Sample Groups

Differential methylation analysis was conducted on site and region level according to the sample groups specified in the analysis.

Comparisons

The following comparisons were made:

P-values

In the following anlyses, p-values on the site level were computed using the limma method. I.e. hierarchical linear models from the limma package were employed and fitted using an empirical Bayes approach on derived M-values.

Site Level

Differential methylation on the site level was computed based on a variety of metrics. Of particular interest for the following plots and analyses are the following quantities for each site: a) the difference in mean methylation levels of the two groups being compared, b) the quotient in mean methylation and c) a statistical test (limma or t-test depending on the settings) assessing whether the methylation values in the two groups originate from distinct distributions. Additionally each site was assigned a rank based on each of these three criteria. A combined rank is computed as the maximum (i.e. worst) rank among the three ranks. The smaller the combined rank for a site, the more evidence for differential methylation it exhibits. This section includes scatterplots of the site group means as well as volcano plots of each pairwise comparison colored according to the combined ranks or p-values of a given site.

The following rank cutfoffs have been automatically selected for the analysis of differentially methylated sites:

Rank Cutoff
Ewing vs. Healthy (based on disease_type) 91956
Ewing_Cell_Line vs. non.Ewing_Cell_Line (based on disease_type_detailed) 71879
Ewing_Tumor vs. non.Ewing_Tumor (based on disease_type_detailed) 75419
MSC_Ewing vs. non.MSC_Ewing (based on disease_type_detailed) 73210
MSC_normal vs. non.MSC_normal (based on disease_type_detailed) 91956
Graz vs. non.Graz (based on sample_source) 331571
Münster vs. non.Münster (based on sample_source) 273050
Paris vs. non.Paris (based on sample_source) 222226
Vienna vs. non.Vienna (based on sample_source) 201492
abdomen vs. non.abdomen (based on tumor_location) 283168
head_and_neck vs. non.head_and_neck (based on tumor_location) 335505
lower_extremity vs. non.lower_extremity (based on tumor_location) 231115
pelvis vs. non.pelvis (based on tumor_location) 244007
spine vs. non.spine (based on tumor_location) 255517
thorax vs. non.thorax (based on tumor_location) 228900
upper_extremity vs. non.upper_extremity (based on tumor_location) 249541
greater_than_200mL vs. less_than_200mL (based on tumor_size) 409269
no_relapse vs. relapse (based on relapse) 156026
metastases vs. no_metastases (based on metastases_at_diagnosis) 164839
Ewing_Cell_Line vs. Ewing_Tumor (based on comp.column) 59565
Ewing_Tumor vs. MSC (based on Tissue_vs_MSC) 123532
comparison
differential methylation measure

Figure 1

Figure 1

Scatterplot for differential methylation (sites). If the selected criterion is not rankGradient: The transparency corresponds to point density. If the number of points exceeds 2e+06 then the number of points for density estimation is reduced to that number by random sampling.The1% of the points in the sparsest populated plot regions are drawn explicitly (up to a maximum of 10000 points).Additionally, the colored points represent differentially methylated sites (according to the selected criterion). If the selected criterion is rankGradient: median combined ranks accross hexagonal bins are shown as a gradient according to the color legend.

comparison
difference metric
significance metric

Figure 2

Figure 2

Volcano plot for differential methylation quantified by various metrics. Color scale according to combined ranking.

Differential Methylation Tables

A tabular overview of measures for differential methylation on the site level for the individual comparisons are provided in this section. Below, a brief explanation of the different columns can be found:

The tables for the individual comparisons can be found here:

Differential Variability

Differentially variable sites were computed with diffVar. For more information about the method, have a look at the missMethyl Bioconductor package.[1] This section contains plots and tables describing the results of this test and further analyses of the sites that were selected as differentially variable. Please note that missing methylation values have been imputed with none.

The following rank cutoffs have been automatically selected for the analysis of differentially variable sites:

Rank Cutoff
Ewing vs. Healthy (based on disease_type) 37696
Ewing_Cell_Line vs. non.Ewing_Cell_Line (based on disease_type_detailed) 41266
Ewing_Tumor vs. non.Ewing_Tumor (based on disease_type_detailed) 49972
MSC_Ewing vs. non.MSC_Ewing (based on disease_type_detailed) 95107
MSC_normal vs. non.MSC_normal (based on disease_type_detailed) 37696
Graz vs. non.Graz (based on sample_source) 165164
Münster vs. non.Münster (based on sample_source) 117306
Paris vs. non.Paris (based on sample_source) 91799
Vienna vs. non.Vienna (based on sample_source) 76835
abdomen vs. non.abdomen (based on tumor_location) 40305
head_and_neck vs. non.head_and_neck (based on tumor_location) 84587
lower_extremity vs. non.lower_extremity (based on tumor_location) 86053
pelvis vs. non.pelvis (based on tumor_location) 76683
spine vs. non.spine (based on tumor_location) 97190
thorax vs. non.thorax (based on tumor_location) 81168
upper_extremity vs. non.upper_extremity (based on tumor_location) 42071
greater_than_200mL vs. less_than_200mL (based on tumor_size) 80857
no_relapse vs. relapse (based on relapse) 85233
metastases vs. no_metastases (based on metastases_at_diagnosis) 54706
Ewing_Cell_Line vs. Ewing_Tumor (based on comp.column) 59148
Ewing_Tumor vs. MSC (based on Tissue_vs_MSC) 70237
comparison
differential variability measure

Figure 3

Figure 3

Scatterplot for differential variable sites. The transparency corresponds to point density. If the number of points exceeds 2e+06 then the number of points for density estimation is reduced to that number by random sampling. The1% of the points in the sparsest populated plot regions are drawn explicitly (up to a maximum of 10000 points). Additionally, the colored points represent differentially variable sites (according to the selected criterion).

comparison
significance metric

Figure 4

Figure 4

Volcano plot for differential variable sites.

comparison
rankCutoff

Figure 5

Figure 5

Scatterplot comparing differentially methylated (DMCs) and variable sites (DVCs), as well as sites that are both differentially methylated and variable. The dotted lines corrspond to the respective rank cutoffs used to call a site differentially methylated/variable.

Region Level

Differential methylation on the region level was computed based on a variety of metrics. Of particular interest for the following plots and analyses are the following quantities for each region: the mean difference in means across all sites in a region of the two groups being compared and the mean of quotients in mean methylation as well as a combined p-value calculated from all site p-values in the region [2]. Additionally each region was assigned a rank based on each of these three criteria. A combined rank is computed as the maximum (i.e. worst) value among the three ranks. The smaller the combined rank for a region, the more evidence for differential methylation it exhibits. Regions were defined based on the region types specified in the analysis. This section includes scatterplots of the region group means as well as volcano plots of each pairwise comparison colored according to the combined rank of a given region.

The following rank cutfoffs have been automatically selected for the analysis of differentially methylated regions:

tiling1kb tiling200bp genes promoters cpgislands ensembleRegBuildBPall
Ewing vs. Healthy (based on disease_type) 14628 27135 1516 1506 1660 8924
Ewing_Cell_Line vs. non.Ewing_Cell_Line (based on disease_type_detailed) 13654 21984 3468 3783 2805 7581
Ewing_Tumor vs. non.Ewing_Tumor (based on disease_type_detailed) 20230 38124 1896 2628 3129 7167
MSC_Ewing vs. non.MSC_Ewing (based on disease_type_detailed) 16058 35395 1446 1776 843 8438
MSC_normal vs. non.MSC_normal (based on disease_type_detailed) 14628 27135 1516 1506 1660 8924
Graz vs. non.Graz (based on sample_source) 48701 95930 1194 2150 559 13033
Münster vs. non.Münster (based on sample_source) 5985 21828 116 397 456 1516
Paris vs. non.Paris (based on sample_source) 3906 21131 32 135 250 167
Vienna vs. non.Vienna (based on sample_source) 322 2800 63 51 134 176
abdomen vs. non.abdomen (based on tumor_location) 14835 40170 395 625 555 3839
head_and_neck vs. non.head_and_neck (based on tumor_location) 19429 48273 653 1197 362 4327
lower_extremity vs. non.lower_extremity (based on tumor_location) 248 3527 46 83 73 96
pelvis vs. non.pelvis (based on tumor_location) 560 4485 34 85 80 207
spine vs. non.spine (based on tumor_location) 9122 23494 220 313 85 1679
thorax vs. non.thorax (based on tumor_location) 184 3006 21 24 51 145
upper_extremity vs. non.upper_extremity (based on tumor_location) 3686 12265 161 182 181 766
greater_than_200mL vs. less_than_200mL (based on tumor_size) 1624 12738 156 188 123 431
no_relapse vs. relapse (based on relapse) 540 3372 63 129 84 252
metastases vs. no_metastases (based on metastases_at_diagnosis) 4104 27127 130 118 129 329
Ewing_Cell_Line vs. Ewing_Tumor (based on comp.column) 5226 12430 3188 3212 3108 3330
Ewing_Tumor vs. MSC (based on Tissue_vs_MSC) 32569 63613 1615 964 2025 8194
comparison
regions
differential methylation measure

Figure 6

Figure 6

Scatterplot for differential methylation (regions). If the selected criterion is not rankGradient: The transparency corresponds to point density. The 1% of the points in the sparsest populated plot regions are drawn explicitly. Additionally, the colored points represent differentially methylated regions (according to the selected criterion). If the selected criterion is rankGradient: median combined ranks accross hexagonal bins are shown as a gradient according to the color legend.

comparison
regions
difference metric
significance metric

Figure 7

Figure 7

Volcano plot for differential methylation quantified by various metrics. Color scale according to combined ranking.

Differential Methylation Tables

A tabular overview of measures for differential methylation on the region level for the individual comparisons are provided in this section.

The tables for the individual comparisons can be found here:

tiling1kb tiling200bp genes promoters cpgislands ensembleRegBuildBPall
Ewing vs. Healthy (based on disease_type) csv csv csv csv csv csv
Ewing_Cell_Line vs. non.Ewing_Cell_Line (based on disease_type_detailed) csv csv csv csv csv csv
Ewing_Tumor vs. non.Ewing_Tumor (based on disease_type_detailed) csv csv csv csv csv csv
MSC_Ewing vs. non.MSC_Ewing (based on disease_type_detailed) csv csv csv csv csv csv
MSC_normal vs. non.MSC_normal (based on disease_type_detailed) csv csv csv csv csv csv
Graz vs. non.Graz (based on sample_source) csv csv csv csv csv csv
Münster vs. non.Münster (based on sample_source) csv csv csv csv csv csv
Paris vs. non.Paris (based on sample_source) csv csv csv csv csv csv
Vienna vs. non.Vienna (based on sample_source) csv csv csv csv csv csv
abdomen vs. non.abdomen (based on tumor_location) csv csv csv csv csv csv
head_and_neck vs. non.head_and_neck (based on tumor_location) csv csv csv csv csv csv
lower_extremity vs. non.lower_extremity (based on tumor_location) csv csv csv csv csv csv
pelvis vs. non.pelvis (based on tumor_location) csv csv csv csv csv csv
spine vs. non.spine (based on tumor_location) csv csv csv csv csv csv
thorax vs. non.thorax (based on tumor_location) csv csv csv csv csv csv
upper_extremity vs. non.upper_extremity (based on tumor_location) csv csv csv csv csv csv
greater_than_200mL vs. less_than_200mL (based on tumor_size) csv csv csv csv csv csv
no_relapse vs. relapse (based on relapse) csv csv csv csv csv csv
metastases vs. no_metastases (based on metastases_at_diagnosis) csv csv csv csv csv csv
Ewing_Cell_Line vs. Ewing_Tumor (based on comp.column) csv csv csv csv csv csv
Ewing_Tumor vs. MSC (based on Tissue_vs_MSC) csv csv csv csv csv csv

Differential Variability

Differential variability on the region level was computed similar to differential methylation, but the mean of variances, the log-ratio of the quotient of variances as well as the p-values from the differentiality test were employed. Ranking was performed in line with the ranking of differential methylation.

The following rank cutoffs have been automatically selected for the analysis of differentially variable regions:

tiling1kb tiling200bp genes promoters cpgislands ensembleRegBuildBPall
Ewing vs. Healthy (based on disease_type) 13373 11502 1548 1322 476 6485
Ewing_Cell_Line vs. non.Ewing_Cell_Line (based on disease_type_detailed) 8047 13189 830 1281 1595 2035
Ewing_Tumor vs. non.Ewing_Tumor (based on disease_type_detailed) 10763 19445 952 1195 1018 3128
MSC_Ewing vs. non.MSC_Ewing (based on disease_type_detailed) 18272 27660 1170 2833 2352 7037
MSC_normal vs. non.MSC_normal (based on disease_type_detailed) 13373 11324 1548 1459 604 6473
Graz vs. non.Graz (based on sample_source) 24441 61810 3474 2552 3353 6790
Münster vs. non.Münster (based on sample_source) 14623 27321 2242 2351 1887 3630
Paris vs. non.Paris (based on sample_source) 8541 20263 1075 2483 1152 3415
Vienna vs. non.Vienna (based on sample_source) 6278 18665 1568 1771 1084 6349
abdomen vs. non.abdomen (based on tumor_location) 15556 10277 2213 2911 3246 6112
head_and_neck vs. non.head_and_neck (based on tumor_location) 24311 32741 1970 3794 2479 4857
lower_extremity vs. non.lower_extremity (based on tumor_location) 14380 16144 89 1203 229 2556
pelvis vs. non.pelvis (based on tumor_location) 11406 27821 132 1224 329 3363
spine vs. non.spine (based on tumor_location) 19903 39972 797 1485 967 4524
thorax vs. non.thorax (based on tumor_location) 8378 27482 284 2880 2428 1928
upper_extremity vs. non.upper_extremity (based on tumor_location) 9879 21115 2229 1964 1775 5223
greater_than_200mL vs. less_than_200mL (based on tumor_size) 18493 24864 132 664 427 3631
no_relapse vs. relapse (based on relapse) 11832 20200 96 1558 367 5716
metastases vs. no_metastases (based on metastases_at_diagnosis) 11641 20847 215 1638 949 3524
Ewing_Cell_Line vs. Ewing_Tumor (based on comp.column) 8323 14858 766 1444 869 2949
Ewing_Tumor vs. MSC (based on Tissue_vs_MSC) 12233 26203 1214 2763 1412 4450
comparison
regions
differential variability measure

Figure 8

Figure 8

Scatterplot for differential variable regions. The transparency corresponds to point density. The 1% of the points in the sparsest populated plot regions are drawn explicitly. Additionally, the colored points represent differentially methylated regions (according to the selected criterion).

comparison
regions
difference metric
significance metric

Figure 9

Figure 9

Volcano plot for differential variability quantified by various metrics. Color scale according to combined ranking.

comparison
regions
rankCutoff

Figure 10

Figure 10

Scatterplot comparing differentially methylated (DMRs) and variable regions (DVRs), as well as regions that are both differentially methylated and variable. The dotted lines corrspond to the respective rank cutoffs used to call a region differentially methylated/variable.

GO Enrichment Analysis

GO Enrichment Analysis was conducted. The wordclouds and tables below contains significant GO terms as determined by a hypergeometric test.

comparison
Hypermethylation/hypomethylation
ontology
regions
differential methylation measure

Figure 11

Figure 11

Wordclouds for GO enrichment terms.

comparison
Hypermethylation/hypomethylation
ontology
regions
differential methylation measure

GOMFID Pvalue OddsRatio ExpCount Count Size Term
GO:1903706 4e-04 9.6333 0.6451 5 409 regulation of hemopoiesis
GO:0060216 4e-04 81.1551 0.03 2 19 definitive hemopoiesis
GO:0048534 0.0012 6.0289 1.2649 6 802 hematopoietic or lymphoid organ development
GO:0071345 0.0013 6.0209 1.2665 6 803 cellular response to cytokine stimulus
GO:0010734 0.0016 Inf 0.0016 1 1 negative regulation of protein glutathionylation
GO:1902867 0.0016 Inf 0.0016 1 1 negative regulation of retina development in camera-type eye
GO:1902870 0.0016 Inf 0.0016 1 1 negative regulation of amacrine cell differentiation
GO:0030099 0.0021 8.457 0.5599 4 355 myeloid cell differentiation
GO:0045596 0.0022 6.4116 0.9526 5 604 negative regulation of cell differentiation
GO:0050902 0.0032 660.5217 0.0032 1 2 leukocyte adhesive activation
GO:0071461 0.0032 660.5217 0.0032 1 2 cellular response to redox state
GO:1901003 0.0032 660.5217 0.0032 1 2 negative regulation of fermentation
GO:0051239 0.0042 3.4436 4.1322 10 2620 regulation of multicellular organismal process
GO:0071560 0.0043 10.2919 0.3328 3 211 cellular response to transforming growth factor beta stimulus
GO:0070315 0.0047 330.2391 0.0047 1 3 G1 to G0 transition involved in cell differentiation
GO:1990791 0.0047 330.2391 0.0047 1 3 dorsal root ganglion development
GO:0045652 0.0048 21.4901 0.1041 2 66 regulation of megakaryocyte differentiation
GO:0007275 0.0049 3.1352 7.4191 14 4704 multicellular organism development
GO:0050793 0.0052 3.4841 3.5345 9 2241 regulation of developmental process
GO:0001501 0.0057 6.3771 0.735 4 466 skeletal system development
GO:0050768 0.0057 9.2529 0.3691 3 234 negative regulation of neurogenesis
GO:1903273 0.0063 220.1449 0.0063 1 4 regulation of sodium ion export
GO:1903278 0.0063 220.1449 0.0063 1 4 positive regulation of sodium ion export across plasma membrane
GO:0048869 0.0075 2.9835 6.0343 12 3826 cellular developmental process
GO:0061056 0.0079 165.0978 0.0079 1 5 sclerotome development
GO:0090259 0.0079 165.0978 0.0079 1 5 regulation of retinal ganglion cell axon guidance
GO:0060842 0.0094 132.0696 0.0095 1 6 arterial endothelial cell differentiation
GO:0061074 0.0094 132.0696 0.0095 1 6 regulation of neural retina development
GO:0045664 0.0097 5.4375 0.8564 4 543 regulation of neuron differentiation

Differential Variability

GO enrichment analysis was also performed for differentially variable regions.

comparison
Hypermethylation/hypomethylation
ontology
regions
differential methylation measure

Figure 12

Figure 12

Workclouds for GO enrichment terms (Differential Variability)

comparison
Hypermethylation/hypomethylation
ontology
regions
differential methylation measure

GOMFID Pvalue OddsRatio ExpCount Count Size Term
GO:0006742 0.0016 Inf 0.0016 1 1 NADP catabolic process
GO:0002587 0.0033 632.9583 0.0033 1 2 negative regulation of antigen processing and presentation of peptide antigen via MHC class II
GO:0050911 0.0039 23.932 0.0936 2 57 detection of chemical stimulus involved in sensory perception of smell
GO:0007165 0.0054 3.0223 8.3032 15 5054 signal transduction
GO:0002580 0.0066 210.9583 0.0066 1 4 regulation of antigen processing and presentation of peptide or polysaccharide antigen via MHC class II
GO:0034653 0.0066 210.9583 0.0066 1 4 retinoic acid catabolic process
GO:0072526 0.0066 210.9583 0.0066 1 4 pyridine-containing compound catabolic process
GO:0002583 0.0082 158.2083 0.0082 1 5 regulation of antigen processing and presentation of peptide antigen
GO:1903232 0.0082 158.2083 0.0082 1 5 melanosome assembly
GO:0051453 0.0089 15.4547 0.1429 2 87 regulation of intracellular pH
GO:0016115 0.0098 126.5583 0.0099 1 6 terpenoid catabolic process

LOLA Enrichment Analysis

LOLA Enrichment Analysis [3] was conducted. The plots and tables below show enrichments across annotations in the supplied LOLA reference databases for the following collections:

comparison
Hypermethylation/hypomethylation
regions
differential methylation measure
color

Figure 13

Figure 13

Scatter plot showing the effect size (log-odds ratio) vs. the significance (-log10(q-value)), similar to a 'volcano plot' as it is called in other contexts.

comparison
Hypermethylation/hypomethylation
regions
differential methylation measure

Figure 14

Open PDF Figure 14

Boxplots showing log-odds ratios from LOLA enrichment analysis. Shown are those groups of terms per category that share the same putative target. Only terms that exhibit statistical significance (p-value < 0.01) are included. If more than 100 terms are enriched, the 100 terms receiving the highest joined LOLA ranks are shown. Coloring of the bars reflects the putative targets of the terms.

comparison
Hypermethylation/hypomethylation
regions
differential methylation measure

Figure 15

Open PDF Figure 15

Barplots showing log-odds ratios from LOLA enrichment analysis. Shown are those terms that exhibit statistical significance (p-value < 0.01). If more than 100 terms are enriched, the 100 terms receiving the highest joined LOLA ranks are shown. Coloring of the bars reflects the putative targets of the terms.

Differential Variability

LOLA enrichment analysis was also conducted for differentially variable regions.

comparison
Hypermethylation/hypomethylation
regions
differential methylation measure
color

Figure 16

Figure 16

Barplots showing log-odds ratios from LOLA enrichment analysis. Shown are those terms that exhibit statistical significance (p-value < 0.01). If more than 100 terms are enriched, the 100 terms receiving the highest joined LOLA ranks are shown. Coloring of the bars reflects the putative targets of the terms.

comparison
Hypermethylation/hypomethylation
regions
differential methylation measure

Figure 17

Open PDF Figure 17

Barplots showing log-odds ratios from LOLA enrichment analysis. Shown are those terms that exhibit statistical significance (p-value < 0.01). If more than 100 terms are enriched, the 100 terms receiving the highest joined LOLA ranks are shown. Coloring of the bars reflects the putative targets of the terms.

comparison
Hypermethylation/hypomethylation
regions
differential methylation measure

Figure 18

Open PDF Figure 18

Barplots showing log-odds ratios from LOLA enrichment analysis. Shown are those terms that exhibit statistical significance (p-value < 0.01). If more than 100 terms are enriched, the 100 terms receiving the highest joined LOLA ranks are shown. Coloring of the bars reflects the putative targets of the terms.

References

  1. Phipson, B., & Oshlack, A. (2014). DiffVar: a new method for detecting differential variability with application to methylation in cancer and aging. Genome Biology, 15(9), 465
  2. Makambi, K. (2003) Weighted inverse chi-square method for correlated significance tests. Journal of Applied Statistics, 30(2), 225234
  3. Sheffield, N. C., & Bock, C. (2016). LOLA: enrichment analysis for genomic region sets and regulatory elements in R and Bioconductor. Bioinformatics, 32(4), 587-589