Skip to main content

Table 1 Running ModuleDigger in the presence of a gradual increase in noise.

From: ModuleDigger: an itemset mining framework for the detection of cis-regulatory modules

Number of TFs included

Test set

10 TFs

20 TFs

30 TFs

40 TFs

Runs

RANK

S

FP

RANK

S

FP

RANK

S

FP

RANK

S

FP

1

6

y

1

18

y

3

41

y

29

15

y

12

2

12

y

2

9

y

2

35

y

3

153

y

2

3

12

y

2

11

y

3

45

y

39

55

y

1

4

11

y

3

14

y

14

71

y

66

54

y

9

5

3

y

1

12

y

1

15

n

18

138

y

52

6

1

y

0

8

y

3

38

y

1

48

y

23

7

1

y

0

4

y

3

52

y

3

141

y

49

8

2

y

0

6

y

3

58

y

1

147

y

3

9

4

y

2

12

y

11

32

y

1

155

y

52

10

1

y

0

5

y

3

45

y

39

158

y

81

Average

6.3

/

1.1

10

/

4.5

43.2

/

20

108

/

28.4

Median

5

/

1

10

/

3

43

/

10.5

139

/

17.5

Std

4.7

/

1.1

4.3

/

4

15.2

/

22.6

56

/

28.0

  1. For each specified number of TFs, 10 different runs were performed which differed in the PWMs randomly selected from TRANSFAC. Average, Median and Std: average, median and standard deviation of the rank of the biologically valid module (OCT4, SOX2 and NANOG). RANK: the rank the valid module received in our output; Score of valid module versus random modules (S): assesses whether the score of the true module is higher than the score of an equally ranked module in a randomized dataset (y = yes, n = no). Number of false positives (FP): the number of modules in the randomized dataset that contained a score higher than the score of the valid module in our benchmark dataset.