Table 5 Performance comparison of several prediction tools based on the PhosPlus_set.

From: PhosphoPredict: A bioinformatics tool for prediction of human kinase-specific phosphorylation substrates and sites by integrating heterogeneous feature selection

Kinase

Method

Accuracy

Sensitivity

Specificity

Precision

Recall

F-Score

MCC

AUC

CDKs

KinasePhos

86.6

65.2

86.9

5.8

65.2

10.6

0.195

0.777

PPSP

91.0

74.1

91.2

9.4

74.1

16.8

0.261

0.838

GPS

84.4

78.0

84.5

5.8

78.0

10.9

0.206

0.881

Musite

88.9

77.1

89.0

8.0

77.1

14.4

0.242

0.886

PhosphoPredict

94.2

77.1

94.4

14.5

77.1

24.4

0.330

0.904

CK2

KinasePhos

89.2

51.2

90.0

9.4

51.2

16.0

0.229

0.714

PPSP

93.1

49.4

94.0

14.4

49.4

22.3

0.274

0.838

GPS

94.1

50.0

95.0

17.0

50.0

25.4

0.298

0.821

Musite

96.4

41.6

97.5

25.5

41.6

33.1

0.331

0.809

PhosphoPredict

91.9

50.6

92.8

12.5

50.6

20.1

0.259

0.727

PKA

KinasePhos

90.4

61.6

90.9

11.1

61.6

18.9

0.264

0.775

PPSP

90.2

73.3

90.5

12.5

73.3

21.3

0.298

0.836

GPS

85.3

80.1

85.4

8.9

80.1

16.0

0.256

0.880

Musite

88.9

70.4

89.2

10.8

70.4

18.7

0.273

0.877

PhosphoPredict

91.1

80.5

91.3

14.0

80.5

32.7

0.327

0.896

PKC

KinasePhos

81.8

49.4

82.3

4.0

49.4

7.4

0.155

0.677

PPSP

83.8

58.8

84.2

5.3

58.8

9.7

0.183

0.734

GPS

82.1

56.8

82.7

6.6

56.8

11.8

0.203

0.785

Musite

86.7

52.3

87.2

5.8

52.3

10.4

0.183

0.798

PhosphoPredict

87.8

57.2

88.3

6.8

57.2

12.2

0.203

0.826

  1. The best results for each kinase and performance measure are highlighted in bold.