Correlation between gene methylation status and clinical features
Skin Cutaneous Melanoma (Metastatic)
23 September 2013  |  analyses__2013_09_23
Maintainer Information
Citation Information
doi:10.7908/C1FJ2F50
Maintained by Juok Cho (Broad Institute)
Cite as Broad Institute TCGA Genome Data Analysis Center (2013): Correlation between gene methylation status and clinical features. Broad Institute of MIT and Harvard. doi:10.7908/C1FJ2F50
Overview
Introduction

This pipeline uses various statistical tests to identify genes whose promoter methylation levels correlated to selected clinical features.

Summary

Testing the association between 19704 genes and 11 clinical features across 239 samples, statistically thresholded by Q value < 0.05, 9 clinical features related to at least one genes.

  • 1 gene correlated to 'Time to Death'.

    • OSBPL2

  • 11 genes correlated to 'AGE'.

    • STL ,  NFATC2 ,  TES ,  TRPV4 ,  SP1 ,  ...

  • 17 genes correlated to 'PRIMARY.SITE.OF.DISEASE'.

    • TMEM101 ,  ACTN3 ,  ZDHHC24__1 ,  EXOC3 ,  EPHB4 ,  ...

  • 12 genes correlated to 'NEOPLASM.DISEASESTAGE'.

    • AGPAT9 ,  C18ORF22 ,  C5ORF43 ,  FBXO30 ,  B3GNT2 ,  ...

  • 1 gene correlated to 'PATHOLOGY.T.STAGE'.

    • FAM100B

  • 284 genes correlated to 'PATHOLOGY.M.STAGE'.

    • LMF1 ,  C10ORF88 ,  FAM186A ,  FGFR2 ,  FRMD5 ,  ...

  • 34 genes correlated to 'MELANOMA.PRIMARY.KNOWN'.

    • ZBTB8A ,  SERPINB1 ,  CARD11 ,  CMTM7 ,  NMUR1 ,  ...

  • 1 gene correlated to 'BRESLOW.THICKNESS'.

    • FAM100B

  • 1 gene correlated to 'GENDER'.

    • DDX43

  • No genes correlated to 'PATHOLOGY.N.STAGE', and 'MELANOMA.ULCERATION'.

Results
Overview of the results

Complete statistical result table is provided in Supplement Table 1

Table 1.  Get Full Table This table shows the clinical features, statistical methods used, and the number of genes that are significantly associated with each clinical feature at Q value < 0.05.

Clinical feature Statistical test Significant genes Associated with                 Associated with
Time to Death Cox regression test N=1 shorter survival N=0 longer survival N=1
AGE Spearman correlation test N=11 older N=11 younger N=0
PRIMARY SITE OF DISEASE ANOVA test N=17        
NEOPLASM DISEASESTAGE ANOVA test N=12        
PATHOLOGY T STAGE Spearman correlation test N=1 higher stage N=1 lower stage N=0
PATHOLOGY N STAGE Spearman correlation test   N=0        
PATHOLOGY M STAGE ANOVA test N=284        
MELANOMA ULCERATION t test   N=0        
MELANOMA PRIMARY KNOWN t test N=34 yes N=32 no N=2
BRESLOW THICKNESS Spearman correlation test N=1 higher breslow.thickness N=1 lower breslow.thickness N=0
GENDER t test N=1 male N=0 female N=1
Clinical variable #1: 'Time to Death'

One gene related to 'Time to Death'.

Table S1.  Basic characteristics of clinical feature: 'Time to Death'

Time to Death Duration (Months) 0.2-357.4 (median=48.9)
  censored N = 121
  death N = 112
     
  Significant markers N = 1
  associated with shorter survival 0
  associated with longer survival 1
List of one gene significantly associated with 'Time to Death' by Cox regression test

Table S2.  Get Full Table List of one gene significantly associated with 'Time to Death' by Cox regression test

HazardRatio Wald_P Q C_index
OSBPL2 0.2 1.299e-06 0.026 0.356

Figure S1.  Get High-res Image As an example, this figure shows the association of OSBPL2 to 'Time to Death'. four curves present the cumulative survival rates of 4 quartile subsets of patients. P value = 1.3e-06 with univariate Cox regression analysis using continuous log-2 expression values.

Clinical variable #2: 'AGE'

11 genes related to 'AGE'.

Table S3.  Basic characteristics of clinical feature: 'AGE'

AGE Mean (SD) 55.4 (16)
  Significant markers N = 11
  pos. correlated 11
  neg. correlated 0
List of top 10 genes significantly correlated to 'AGE' by Spearman correlation test

Table S4.  Get Full Table List of top 10 genes significantly correlated to 'AGE' by Spearman correlation test

SpearmanCorr corrP Q
STL 0.3459 5.254e-08 0.00104
NFATC2 0.3322 1.85e-07 0.00364
TES 0.3305 2.152e-07 0.00424
TRPV4 0.3282 2.643e-07 0.00521
SP1 0.3122 1.043e-06 0.0206
ITGA9 0.309 1.36e-06 0.0268
STK38L 0.3064 1.685e-06 0.0332
LRRC8C 0.3042 2.025e-06 0.0399
FBXL13__2 0.3038 2.089e-06 0.0411
LRRC17 0.3038 2.089e-06 0.0411

Figure S2.  Get High-res Image As an example, this figure shows the association of STL to 'AGE'. P value = 5.25e-08 with Spearman correlation analysis. The straight line presents the best linear regression.

Clinical variable #3: 'PRIMARY.SITE.OF.DISEASE'

17 genes related to 'PRIMARY.SITE.OF.DISEASE'.

Table S5.  Basic characteristics of clinical feature: 'PRIMARY.SITE.OF.DISEASE'

PRIMARY.SITE.OF.DISEASE Labels N
  DISTANT METASTASIS 32
  PRIMARY TUMOR 1
  REGIONAL CUTANEOUS OR SUBCUTANEOUS TISSUE 50
  REGIONAL LYMPH NODE 155
     
  Significant markers N = 17
List of top 10 genes differentially expressed by 'PRIMARY.SITE.OF.DISEASE'

Table S6.  Get Full Table List of top 10 genes differentially expressed by 'PRIMARY.SITE.OF.DISEASE'

ANOVA_P Q
TMEM101 8.832e-54 1.74e-49
ACTN3 3.278e-22 6.46e-18
ZDHHC24__1 3.278e-22 6.46e-18
EXOC3 5.453e-15 1.07e-10
EPHB4 3.454e-10 6.8e-06
DKFZP686I15217 5.355e-10 1.05e-05
TUBB3 7.131e-10 1.4e-05
GALR2 1.203e-08 0.000237
NACC2 8.73e-08 0.00172
RHOJ 8.879e-08 0.00175

Figure S3.  Get High-res Image As an example, this figure shows the association of TMEM101 to 'PRIMARY.SITE.OF.DISEASE'. P value = 8.83e-54 with ANOVA analysis.

Clinical variable #4: 'NEOPLASM.DISEASESTAGE'

12 genes related to 'NEOPLASM.DISEASESTAGE'.

Table S7.  Basic characteristics of clinical feature: 'NEOPLASM.DISEASESTAGE'

NEOPLASM.DISEASESTAGE Labels N
  I OR II NOS 9
  STAGE 0 4
  STAGE I 20
  STAGE IA 10
  STAGE IB 22
  STAGE II 17
  STAGE IIA 10
  STAGE IIB 12
  STAGE IIC 8
  STAGE III 20
  STAGE IIIA 10
  STAGE IIIB 21
  STAGE IIIC 37
  STAGE IV 10
     
  Significant markers N = 12
List of top 10 genes differentially expressed by 'NEOPLASM.DISEASESTAGE'

Table S8.  Get Full Table List of top 10 genes differentially expressed by 'NEOPLASM.DISEASESTAGE'

ANOVA_P Q
AGPAT9 5.342e-08 0.00105
C18ORF22 6.704e-08 0.00132
C5ORF43 2.314e-07 0.00456
FBXO30 2.755e-07 0.00543
B3GNT2 4.039e-07 0.00796
N4BP3 5.901e-07 0.0116
LOC728392 6.144e-07 0.0121
DYSFIP1 8.69e-07 0.0171
ELP2 9.43e-07 0.0186
SLC39A6 9.43e-07 0.0186

Figure S4.  Get High-res Image As an example, this figure shows the association of AGPAT9 to 'NEOPLASM.DISEASESTAGE'. P value = 5.34e-08 with ANOVA analysis.

Clinical variable #5: 'PATHOLOGY.T.STAGE'

One gene related to 'PATHOLOGY.T.STAGE'.

Table S9.  Basic characteristics of clinical feature: 'PATHOLOGY.T.STAGE'

PATHOLOGY.T.STAGE Mean (SD) 2.47 (1.2)
  N
  0 14
  1 27
  2 53
  3 46
  4 49
     
  Significant markers N = 1
  pos. correlated 1
  neg. correlated 0
List of one gene significantly correlated to 'PATHOLOGY.T.STAGE' by Spearman correlation test

Table S10.  Get Full Table List of one gene significantly correlated to 'PATHOLOGY.T.STAGE' by Spearman correlation test

SpearmanCorr corrP Q
FAM100B 0.3604 3.506e-07 0.00691

Figure S5.  Get High-res Image As an example, this figure shows the association of FAM100B to 'PATHOLOGY.T.STAGE'. P value = 3.51e-07 with Spearman correlation analysis.

Clinical variable #6: 'PATHOLOGY.N.STAGE'

No gene related to 'PATHOLOGY.N.STAGE'.

Table S11.  Basic characteristics of clinical feature: 'PATHOLOGY.N.STAGE'

PATHOLOGY.N.STAGE Mean (SD) 0.8 (1.1)
  N
  0 126
  1 38
  2 28
  3 27
     
  Significant markers N = 0
Clinical variable #7: 'PATHOLOGY.M.STAGE'

284 genes related to 'PATHOLOGY.M.STAGE'.

Table S12.  Basic characteristics of clinical feature: 'PATHOLOGY.M.STAGE'

PATHOLOGY.M.STAGE Labels N
  M0 210
  M1 3
  M1A 2
  M1B 2
  M1C 5
     
  Significant markers N = 284
List of top 10 genes differentially expressed by 'PATHOLOGY.M.STAGE'

Table S13.  Get Full Table List of top 10 genes differentially expressed by 'PATHOLOGY.M.STAGE'

ANOVA_P Q
LMF1 3.151e-30 6.21e-26
C10ORF88 8.215e-28 1.62e-23
FAM186A 4.608e-27 9.08e-23
FGFR2 3.585e-24 7.06e-20
FRMD5 4.591e-24 9.04e-20
LOC728758 4.591e-24 9.04e-20
HSPA9 6.204e-21 1.22e-16
SNORD63 6.204e-21 1.22e-16
COL5A1 6.607e-21 1.3e-16
FAM186B 1.003e-20 1.97e-16

Figure S6.  Get High-res Image As an example, this figure shows the association of LMF1 to 'PATHOLOGY.M.STAGE'. P value = 3.15e-30 with ANOVA analysis.

Clinical variable #8: 'MELANOMA.ULCERATION'

No gene related to 'MELANOMA.ULCERATION'.

Table S14.  Basic characteristics of clinical feature: 'MELANOMA.ULCERATION'

MELANOMA.ULCERATION Labels N
  NO 91
  YES 59
     
  Significant markers N = 0
Clinical variable #9: 'MELANOMA.PRIMARY.KNOWN'

34 genes related to 'MELANOMA.PRIMARY.KNOWN'.

Table S15.  Basic characteristics of clinical feature: 'MELANOMA.PRIMARY.KNOWN'

MELANOMA.PRIMARY.KNOWN Labels N
  NO 30
  YES 209
     
  Significant markers N = 34
  Higher in YES 32
  Higher in NO 2
List of top 10 genes differentially expressed by 'MELANOMA.PRIMARY.KNOWN'

Table S16.  Get Full Table List of top 10 genes differentially expressed by 'MELANOMA.PRIMARY.KNOWN'

T(pos if higher in 'YES') ttestP Q AUC
ZBTB8A 6.21 2.514e-09 4.95e-05 0.6338
SERPINB1 6.08 1.41e-08 0.000278 0.6896
CARD11 5.86 1.679e-08 0.000331 0.6579
CMTM7 5.81 2.173e-08 0.000428 0.6459
NMUR1 5.71 3.435e-08 0.000677 0.6126
UBA7 5.72 3.772e-08 0.000743 0.656
NPAS3 5.79 4.291e-08 0.000845 0.6022
MIR564 5.59 6.309e-08 0.00124 0.5978
TMEM42 5.59 6.309e-08 0.00124 0.5978
TRA2A 5.59 6.985e-08 0.00138 0.5861

Figure S7.  Get High-res Image As an example, this figure shows the association of ZBTB8A to 'MELANOMA.PRIMARY.KNOWN'. P value = 2.51e-09 with T-test analysis.

Clinical variable #10: 'BRESLOW.THICKNESS'

One gene related to 'BRESLOW.THICKNESS'.

Table S17.  Basic characteristics of clinical feature: 'BRESLOW.THICKNESS'

BRESLOW.THICKNESS Mean (SD) 3.59 (5.2)
  Significant markers N = 1
  pos. correlated 1
  neg. correlated 0
List of one gene significantly correlated to 'BRESLOW.THICKNESS' by Spearman correlation test

Table S18.  Get Full Table List of one gene significantly correlated to 'BRESLOW.THICKNESS' by Spearman correlation test

SpearmanCorr corrP Q
FAM100B 0.3783 1.919e-07 0.00378

Figure S8.  Get High-res Image As an example, this figure shows the association of FAM100B to 'BRESLOW.THICKNESS'. P value = 1.92e-07 with Spearman correlation analysis. The straight line presents the best linear regression.

Clinical variable #11: 'GENDER'

One gene related to 'GENDER'.

Table S19.  Basic characteristics of clinical feature: 'GENDER'

GENDER Labels N
  FEMALE 93
  MALE 146
     
  Significant markers N = 1
  Higher in MALE 0
  Higher in FEMALE 1
List of one gene differentially expressed by 'GENDER'

Table S20.  Get Full Table List of one gene differentially expressed by 'GENDER'

T(pos if higher in 'MALE') ttestP Q AUC
DDX43 -5.15 5.602e-07 0.011 0.712

Figure S9.  Get High-res Image As an example, this figure shows the association of DDX43 to 'GENDER'. P value = 5.6e-07 with T-test analysis.

Methods & Data
Input

  • Expresson data file = SKCM-TM.meth.by_min_expr_corr.data.txt

  • Clinical data file = SKCM-TM.clin.merged.picked.txt

  • Number of patients = 239

  • Number of genes = 19704

  • Number of clinical features = 11

Survival analysis

For survival clinical features, Wald's test in univariate Cox regression analysis with proportional hazards model (Andersen and Gill 1982) was used to estimate the P values using the 'coxph' function in R. Kaplan-Meier survival curves were plot using the four quartile subgroups of patients based on expression levels

Correlation analysis

For continuous numerical clinical features, Spearman's rank correlation coefficients (Spearman 1904) and two-tailed P values were estimated using 'cor.test' function in R

ANOVA analysis

For multi-class clinical features (ordinal or nominal), one-way analysis of variance (Howell 2002) was applied to compare the log2-expression levels between different clinical classes using 'anova' function in R

Student's t-test analysis

For two-class clinical features, two-tailed Student's t test with unequal variance (Lehmann and Romano 2005) was applied to compare the log2-expression levels between the two clinical classes using 't.test' function in R

Q value calculation

For multiple hypothesis correction, Q value is the False Discovery Rate (FDR) analogue of the P value (Benjamini and Hochberg 1995), defined as the minimum FDR at which the test may be called significant. We used the 'Benjamini and Hochberg' method of 'p.adjust' function in R to convert P values into Q values.

Download Results

In addition to the links below, the full results of the analysis summarized in this report can also be downloaded programmatically using firehose_get, or interactively from either the Broad GDAC website or TCGA Data Coordination Center Portal.

References
[1] Andersen and Gill, Cox's regression model for counting processes, a large sample study, Annals of Statistics 10(4):1100-1120 (1982)
[2] Spearman, C, The proof and measurement of association between two things, Amer. J. Psychol 15:72-101 (1904)
[3] Howell, D, Statistical Methods for Psychology. (5th ed.), Duxbury Press:324-5 (2002)
[4] Lehmann and Romano, Testing Statistical Hypotheses (3E ed.), New York: Springer. ISBN 0387988645 (2005)
[5] Benjamini and Hochberg, Controlling the false discovery rate: a practical and powerful approach to multiple testing, Journal of the Royal Statistical Society Series B 59:289-300 (1995)
Copyright © 2013 Broad Institute TCGA GDAC as part of the TCGA Research Network. All rights reserved.
Made with Nozzle