Skip to main content

Table 6 Results of LOPO cross validation of our method with respect to 14 RBPs

From: Predicting protein-binding regions in RNA using nucleotide profiles and compositions

 

TP

TN

FP

FN

Sensitivity

Specificity

Accuracy

PPV

NPV

MCC

AGO1

37

50

3

18

67.27%

94.34%

80.56%

92.50%

73.53%

0.638

AGO2

39

49

2

18

68.42%

96.08%

81.48%

95.12%

73.13%

0.664

EWSR1

200

198

14

14

93.46%

93.40%

93.43%

93.46%

93.40%

0.869

FUS

468

534

46

19

96.10%

92.07%

93.91%

91.05%

96.56%

0.879

FXR1

3

7

0

1

75.00%

100.00%

90.91%

100.00%

87.50%

0.810

FXR2

25

33

1

11

69.44%

97.06%

82.86%

96.15%

75.00%

0.688

IGF2BP2

57

55

7

15

79.17%

88.71%

83.58%

89.06%

78.57%

0.678

LIN28A

221

263

25

57

79.50%

91.32%

85.51%

89.84%

82.19%

0.714

LIN28B

2214

2343

329

227

90.70%

87.69%

89.13%

87.06%

91.17%

0.783

QKI

3

5

0

1

75.00%

100.00%

88.89%

100.00%

83.33%

0.791

TAF15

11

16

1

2

84.62%

94.12%

90.00%

91.67%

88.89%

0.796

TARDBP

39

159

14

149

20.74%

91.91%

54.85%

73.58%

51.62%

0.179

YTHDF2

35

39

5

6

85.37%

88.64%

87.06%

87.50%

86.67%

0.741

ZC3H7B

388

438

43

94

80.50%

91.06%

85.77%

90.02%

82.33%

0.720

Total

3,740

4,189

490

632

      

Weighted average

    

85.54%

89.53%

87.60%

88.42%

86.89%

0.752

  1. The weighted average was computed from the total values of TP, TN, FP and FN of all runs. TP: true positive, TN true negative, FP false positive, FN false negative, PPV positive prediction value, NPV negative prediction value, MCC Matthews correlation coefficient