Skip to main content

Table 2 Comparison with existing methods for predicting DNA-binding residues.

From: Predicting DNA-binding locations and orientation on proteins using knowledge-based learning of geometric properties

  

Precision

   

Sensitivity

 

PDB ID: chain ID

NAPS

DISPLAR

The proposed method

 

NAPS

DISPLAR

The proposed method

1KU7:A

0.27

0.55

0.56

 

0.57

0.79

0.71

1RIO:A

0.20

0.81

0.25

 

0.40

0.65

0.30

2R1J:L

0.48

0.84

0.39

 

0.62

0.76

0.43

2Z3X:A

0.35

0.00 a

0.48

 

0.38

0.00

0.57

1A3Q:A

0.08

0.47

0.30

 

0.53

0.53

0.47

3DFX:A

0.24

0.43

0.55

 

0.45

0.65

0.30

2O49:A

0.15

0.38

0.43

 

0.36

0.36

0.86

2E1C:A

0.09

0.88

0.06

 

0.46

0.54

0.15

3ERE:D

0.16

0.67

0.35

 

0.50

0.80

0.30

1BDT:A

0.36

0.00 a

0.30

 

0.57

0.00

0.57

3CLC:A

0.38

0.91

0.75

 

0.56

0.63

0.38

Average

0.25

0.54

0.40

 

0.49

0.52

0.46

  1. The residues discovered by MAGIIC-PRO and with RSA>25% are collected as the prediction set of the proposed method for comparison. The best records among the three methods are bolded for both precision and sensitivity rates in each row.
  2. a no DNA-binding residues were predicted.