Skip to main content

Table 2 Average number of shared differentially expressed genes found by different methods for the CMAP dataset (FDR ≤ 10%)

From: Influence of batch effect correction methods on drug induced differential gene expression profiles

Method

limma+Null

limma+1PC

limma+2PCs

limma+3PCs

limma+4PCs

limma+Batch id

LEAPP

Limma+Null

(ASS = 50)

44 (37.3%)

27 (20.0%)

18 (14.9%)

14 (14.1%)

49 (19.6%)

9 (11.0%)

limma+1PC

44 (88.0%)

(ASS = 118)

61 (45.2%)

39 (32.2%)

28 (18.2%)

106 (42.4%)

15 (18.3%)

limma+2PCs

27 (54.0%)

61 (51.7%)

(ASS = 135)

81 (66.9%)

55 (55.6%)

120 (48.0%)

17 (20.7%)

limma+3PCs

18 (36.0%)

39 (33.1%)

81 (60.0%)

(ASS = 121)

70 (70.7%)

100 (40.0%)

17 (20.7%)

limma+4PCs

14 (28.0%)

28 (23.7%)

55 (40.7%)

70 (57.9%)

(ASS = 99)

75 (30.0%)

15 (18.3%)

limma+Batch id

49 (98.0%)

106 (89.8%)

120 (88.9%)

100 (82.6%)

75 (75.8%)

(ASS = 250)

25 (30.5%)

LEAPP

9 (18.0%)

15 (12.7%)

17 (12.6%)

17 (14.0%)

15 (15.2%)

25 (10.0%)

(ASS = 82)

  1. Abbreviations: ASS = Average signature size (removed missing values)
  2. Note: The table contains the number of differentially expressed genes that are shared between each pair of methods on the CMAP dataset. The numbers on the diagonal indicate the average number of differentially expressed genes found by the respective methods. For the LEAPP method, the significant genes with estimate = 0 were ignored. Percentages in parentheses are the proportions of the number of shared genes to average signature size produced by the method on the column header