Skip to main content

Table 1 Calculation of FDRs for the sample pair SP/RP

From: An assessment of false discovery rates and statistical significance in label-free quantitative proteomics with combined filters

  

Statistical test option

Proteins

Source of significant proteins

t-test

ranksum

either

both

neither

UL

RA 1

32

42

44

30

92

 

RA 2

31

38

41

28

99

 

RA 3

29

38

41

26

107

 

RA 4

28

40

41

27

108

 

MPSP

22

29

29

22

61

IS

RA 1

9

14

15

8

72

 

RA 2

4

10

10

4

57

 

RA 3

5

14

14

5

84

 

RA 4

5

14

14

5

68

 

MPSP

1

5

5

1

28

FDR

0.045

0.172

0.172

0.045

0.459

  1. FDRs were calculated for the sample pair SP/RP based on a separate quantitation of the UL protein relative abundance and the IS protein relative abundance. The fold-change was fixed at 2 and the MPSP at 4. The t-test and the Wilcoxon ranksum test were performed at a 5% significance level. RA1, RA2, RA3, and RA4 represent the protein relative abundance for the four permuted sample-pairings. Shown in a row of 'RAx' (x = 1, 2, 3, or 4 permuted sample-pairings) are the numbers of proteins found significant by the five statistical test options, including: the t-test alone ('t-test'), the ranksum test alone ('ranksum'), either the t-test or the ranksum test ('either'), both the t-test and the ranksum test ('both'), and without a statistical test ('neither'). Shown in the two 'MPSP' rows are the UL and IS proteins, respectively, found significant in all four permuted sample-pairings. A FDR is the ratio of the number of significant IS proteins over the number of significant UL proteins.