Skip to main content

Table 5 Comparison of prediction methods on the PRI727 and PRI267 datasets.

From: Prediction of RNA-binding amino acids from protein and RNA sequences

Approach

Sensitivity (%)

Specificity (%)

Accuracy (%)

NP (%)

Fm (%)

CC

PRI727 dataset

RNABindR (opt)

51.5

90.4

88.0

70.9

65.6

0.31

RNABindR (sp)

29.6

96.6

92.4

63.1

45.3

0.29

RNABindR (sn)

90.2

45.9

48.6

68.0

60.8

0.18

BindN (sn80)

88.4

51.0

53.4

69.7

64.7

0.19

BindN (sp80)

67.2

77.2

76.5

72.2

71.8

0.25

Our method 1

87.2

81.7

82.1

84.5

84.4

0.40

Our method 2

82.1

76.4

76.8

79.2

79.1

0.32

PRI267 dataset

RNABindR (opt)

4.0

98.4

96.4

51.2

7.6

0.03

RNABindR (sp)

0.8

99.8

97.8

50.3

1.7

0.02

RNABindR (sn)

65.9

57.9

58.1

61.9

61.7

0.07

BindN (sn80)

80.2

53.5

54.0

66.8

64.2

0.10

BindN (sp80)

45.6

80.2

79.5

62.9

58.1

0.09

Our method 1

60.7

91.0

90.3

75.8

72.8

0.24

Our method 2

48.0

85.6

84.8

66.8

61.5

0.13

  1. ’Our method 1’ used all the 9 features (2 global features and 6 local features of protein and the RNA feature), whereas ’Our method 2’ used the 8 protein features (2 global features and 6 local features of protein) only. sn: high sensitivity option. sp: high specificity option. opt: optimal option. sn80: expected sensitivity of 80%. sp80: expected specificity of 80%. NP: net prediction. Fm: F-measure. CC: correlation coefficient.