Skip to main content

Table 3 Validation of genes with differential DNA methylation as predictors of hormone status from the Illumina BCCC (training dataset) using TCGA (validation dataset)

From: DNA methylation and hormone receptor status in breast cancer

 

Training dataset

Validation dataset

Correlation with expression

Gene

Associationa

SAM d scoreb

Associationa

SAM d scorec

ρ d

P valued

FZD9

Positive

3.94

Positive

8.65

−0.44

1.4E−15

MME

Positive

2.95

Positive

2.14

−0.13

2.1E−02

RAB32

Positive

2.70

Positivee

1.11

−0.27

1.4E−06

BCAP31

Positive

2.66

Positive

5.61

−0.28

4.9E−07

HDAC9

Positive

2.64

Positive

7.60

−0.15

8.5E−03

PAX6

Positive

2.64

Positive

4.56

−0.27

2.0E−06

SCGB3A1

Positive

2.53

Positive

9.51

−0.29

3.8E−07

PDGFRA

Positive

2.52

Positive

2.09

−0.30

1.1E−07

IGFBP3

Positive

2.51

Positive

6.37

−0.22

1.1E−04

PTGS2

Positive

2.50

Positive

5.69

−0.30

5.2E−08

SRC

Positive

2.50

Not-associated

0.00

NA

NA

CHI3L2

Positive

2.45

Positive

2.65

−0.69

2.2E−44

PGR

Positive

2.44

Positive

5.39

0.34

1.3E−09

TMPRSS4

Positive

2.43

NA

NA

NA

NA

RASSF1

Positive

2.43

Positive

7.78

−0.05

4.2E−01

TBX1

Positive

2.43

Positive

4.62

−0.05

4.2E−01

PARP1

Positive

2.38

Positive

2.48

−0.12

2.0E−02

COL1A1

Positive

2.32

Positive

4.15

0.08

1.7E−01

SOX17

Positive

2.32

Positive

2.22

−0.13

5.7E−05

RUNX3

Positive

2.29

Positive

7.06

−0.13

2.0E−02

TES

Positive

2.23

Positive

2.15

−0.45

2.6E−16

GPATC3

Positive

2.21

Positivee

0.17

NA

NA

S100A2

Positive

2.21

Positive

9.32

−0.52

2.6E−22

MYH11

Positive

2.20

Positive

3.61

−0.10

8.0E−02

BMP2

Positive

2.19

Positive

4.66

−0.37

1.3E−11

  1. NA gene was absent in the dataset
  2. aIndicates whether gene hypermethylation was associated with increased likelihood of ER/PR-positive breast cancer versus ER/PR-negative breast cancer (“Positive”)
  3. b d scores from SAM analysis using Δ of 0.7 on the GoldenGate dataset
  4. c d scores from SAM analysis using Δ of 3 on the TCGA dataset. In cases where several probes per gene were present, the data is shown for the probe with the highest SAM d score
  5. dPearson correlation coefficient between methylation and expression from TCGA and the corresponding P value
  6. eNon-significant association