Skip to main content

Table 6 A comparison of the GPP predictor and other glycosylation prediction programs.

From: Prediction of glycosylation sites using random forests

 

GPP

NetOglyc

NetNglyc

Oglyc

CKSAAP [27]

EnsembleGly [26]

Scan Site

Ser CCI

90.8

91.8

N/A

N/R

83.1

N/R

N/A

Ser Sensitivity

96.1

66.7

N/A

N/R

80.7

N/R

N/A

Ser Specificity

88.9

95.3

N/A

N/R

85.6

N/R

N/A

Ser MCC

0.81

0.62

N/A

N/R

0.671

N/R

N/A

Thr CCI

92.0

84.9

N/A

N/R

81.4

N/R

N/A

Thr Sensitivity

93.6

81.5

N/A

N/R

80.3

N/R

N/A

Thr Specificity

92.4

89.5

N/A

N/R

82.5

N/R

N/A

Thr MCC

0.84

0.67

N/A

N/R

0.63

N/R

N/A

Asn CCI

92.8

N/A

76.7

N/A

N/A

95.0

79.8

Asn Sensitivity

96.6

N/A

43.9

N/A

N/A

98.0

72.7

Asn Specificity

91.8

N/A

95.7

N/A

N/A

77.0b

81.9

Asn MCC

0.85

N/A

0.49

N/A

N/A

0.84

0.54

Overall CCI

91.4a

88.6a

N/A

87.0a

N/A

89.0

N/A

Overall Sensitivity

94.9a

76.0a

N/A

92.0a

N/A

59.0

N/A

Overall Specificity

90.7a

92.8a

N/A

78.0a

N/A

68.0b

N/A

Overall MCC

0.83a

0.66a

N/A

0.71a

N/A

0.64

N/A

  1. a. combined accuracy for Ser and Thr
  2. b. Specificity for EnsembleGly was calculated as T p /T p +F p . See text for comparison.
  3. N/A = not applicable; N/R = not reported; CCI = % correctly classified instances; MCC = Matthews Correlation Coefficient