Skip to main content

Table 3 Top component memberships

From: Statistical analysis and significance testing of serial analysis of gene expression data using a Poisson mixture model

Dataset

  

Component assignment

Freq.

p -value

BRAIN

astrocytoma (N = 14 nk = 2 = 454)

   -+------------

14

4.3E-28

   ---+++++++++++

12

2.7E-23

ependymoma (N = 10 nk = 2 = 544)

   -+--------

16

6.0E-14

   --+-------

15

9.3E-13

glioblastoma (N = 7 nk = 2 = 607)

   ---+---

48

3.2E-19

   -----+-

42

5.2E-15

medulloblastoma (N = 18 nk = 2 = 438)

   -----------------+

4

1.5E-9

   -----------+-----+

4

1.5E-9

normal (N = 9 nk = 2 = 588)

   -----+--

41

7.4E-37

   -----+-+

21

7.6E-14

AML

inv(16) (N = 5 nk = 2 = 387)

   -++++

110

2.5E-39

   -+-++

36

0.028

t(8;21) (N = 5 nk = 2 = 387)

   -++--

110

2.5E-39

   -++-+

36

0.028

t(15;17) (N = 5 nk = 2 = 225)

   -++--

110

4.5E-67

   -++-+

36

1.0E-6

t(9;11) de novo (N = 4 nk = 2 = 502)

   -+-+

143

1.6E-16

   -+--

132

1.4E-12

t(9;11) treated (N = 3 nk = 2 = 405)

   -++

216

1.1E-16

BREAST

normal (N = 6 nk = 2 = 571)

   ---+--

82

3.6E-29

   -----+

65

4.0E-18

DCIS (N = 4 nk = 2 = 154)

   --++

97

1.4E-43

   -+++

41

4.5E-5

invasive (N = 3 nk = 2 = 765)

   -+-

337

4.6E-10

SKIN

normal (N = 4 nk = 2 = 500)

   ---+

215

1.8E-54

melanoma (N = 3 nk = 2 = 650)

   -+-

405

2.1E-51

  1. For each set of biological replicates, the top one or two component states were selected from tags where the estimated number of components is 2. One component was represented with -, the other with + (i.e. -+- is equivalent to +-+). The significance was calculated using a zero-truncated Binomial test. The number of possible ways for the libraries to be assigned to the two components is (2N/2)-1, where N is the number of libraries.