Skip to main content

Table 1 Performance of SUMOpre using datasets filtering out highly homologous sequences.

From: A novel method for high accuracy sumoylation site prediction from protein sequences

Dataset

Similaritya

Coverageb

size

Methodc

CC

Sn (%)

Sp (%)

Ac (%)

Dataset 1

0.3

0.4

108

self

0.6782

64.02

99.13

97.80

    

jk

0.6364

60.85

98.94

97.50

Dataset 2

0.3

0.6

119

self

0.6500

66.03

98.67

97.41

    

jk

0.6061

63.64

98.35

97.01

Dataset 3

0.3

0.8

140

self

0.6520

70.34

98.33

97.24

    

jk

0.5983

66.53

97.95

96.72

All data

-

-

159

self

0.6401

74.25

97.74

96.79

   

(268 sites)

jk

0.5911

70.90

97.30

96.23

  1. a Similarity threshold used in NCBI BLASTCLUST was set as 0.3;
  2. b Minimum length coverage threshold used in NCBI BLASTCLUST;
  3. c Validation strategies: self, self-consistency test; jk, jack-knife validation