Skip to main content

Table 2 Consensus set statistics

From: Breast cancer subtype predictors revisited: from consensus to concordance?

Dataset

Chip

Nr. of samples

Nr. of samples (%)

 
  

after QC

CS

Basal

HER2

LumA

LumB

Bos

hgu133plus2

188

119 (63.3)

49 (41.2)

19 (16.0)

23 (19.3)

28 (23.5)

expO

hgu133plus2

333

213 (64.0)

56 (26.3)

20 (9.4)

75 (35.2)

62 (29.1)

Guedj

hgu133plus2

501

235 (46.9)

40 (17.0)

21 (8.9)

88 (37.4)

86 (36.6)

Li

hgu133plus2

109

83 (76.1)

25 (30.1)

10 (12.0)

29 (34.9)

19 (22.9)

Sabatier

hgu133plus2

242

162 (66.9)

63 (38.9)

15 (9.3)

40 (24.7)

44 (27.2)

Total

 

1373

812 (59.1)

233 (28.7)

85 (10.5)

255 (31.4)

239 (29.4)

  1. Overview of the five training sets (see also Table 1) used for consensus set construction and the resulting consensus sets. Numbers in parentheses represent percentages. For CS, percentages were calculated w.r.t. the number of samples after QC; for the subtypes w.r.t. the size of the CS. The complete set of 812 consensus set samples, including subtype assignments, is available as Additional file 3