Skip to main content

Table 1 Area under ROC curves up to 100 false positives, RMA and GCRMA

From: Empirical Bayes models for multiple probe type microarrays at the probe level

Method

Pre-processing

Affymetrix U95

Affymetrix 133A

Golden Spike

Gene Logic Tonsil

Gene Logic AML

PLW

RMA

96(1)

93(6)

42(1)

87(1)

86(1)

LMW

RMA

96(2)

94(1)

36(5)

84(3)

80(5)

LPE

RMA

94(7)

93(11)

40(2)

84(2)

85(2)

combined-p

RMA

95(4)

92(12)

39(3)

83(4)

81(4)

WAME

RMA

95(5)

94(2)

33(7)

81(7)

78(8)

median-t

RMA

95(3)

93(10)

39(4)

82(6)

80(6)

IBMT

RMA

95(6)

94(3)

34(6)

78(9)

76(9)

Efron-t

RMA

94(8)

93(4)

32(8)

79(8)

79(7)

FC

RMA

92(12)

93(5)

29(12)

83(5)

85(3)

LIMMA

RMA

94(9)

93(7)

32(9)

76(10)

75(10)

SAM

RMA

94(10)

93(8)

32(11)

74(12)

74(11)

Shrink-t

RMA

94(11)

93(9)

32(10)

75(11)

73(12)

t-test

RMA

85(13)

86(13)

21(13)

57(13)

52(13)

PLW

GCRMA

97(1)

92(8)

54(1)

87(1)

87(1)

LMW

GCRMA

95(3)

93(1)

50(5)

84(3)

79(6)

median-t

GCRMA

96(2)

92(10)

50(2)

83(4)

81(5)

combined-p

GCRMA

95(5)

91(12)

50(4)

86(2)

81(4)

LPE

GCRMA

95(6)

91(11)

50(3)

82(6)

86(2)

IBMT

GCRMA

95(4)

93(2)

47(6)

81(9)

76(8)

Efron-t

GCRMA

94(7)

93(4)

37(8)

82(7)

79(7)

WAME

GCRMA

94(10)

93(3)

39(7)

83(5)

75(10)

FC

GCRMA

93(12)

93(7)

30(13)

81(8)

86(3)

LIMMA

GCRMA

94(9)

93(5)

37(9)

80(10)

73(11)

SAM

GCRMA

94(8)

93(6)

36(11)

79(11)

76(9)

Shrink-t

GCRMA

94(11)

92(9)

37(10)

78(12)

70(12)

t-test

GCRMA

86(13)

84(13)

30(12)

64(13)

53(13)

  1. Area under ROC curves up to 100 false positives rounded to nearest integer value with an optimum of 100. Numbers within parenthesis are within data set ranks for the methods compared (separately for RMA and GCRMA). Methods are ordered with respect to mean rank across the five data sets. Results in the upper and lower part are based on RMA and GCRMA pre-processed data, respectively.