Skip to main content

Table 3 DeepFrag-k and ProFold folds classifications accuracies for DD-dataset

From: DeepFrag-k: a fragment-based deep learning approach for protein fold recognition

#

Fold ID

Fold Name

DeepFrag-k Accuracy

ProFold Accuracy

1

a.1

Globin-like

98.0

100.0

2

a.3

Cytochrome c

95.0

100.0

3

a.4

DNA/RNA-binding 3-helical bundle

85.9

60.0

4

a.24

4-Helical up-and-down bundle

91.5

87.5

5

a.26

4-Helical cytokines

98.9

88.9

6

a.39

EF hand-like

90.8

77.8

7

b.1

Immunoglobulin-like β-sandwich

91.1

84.1

8

b.6

Cupredoxin-like

78.7

66.7

9

b.121

Nucleoplasmin-like/VP

91.3

92.3

10

b.29

ConA-like lectins/glucanases

76.7

66.7

11

b.34

SH3-like barrel

78.0

50.0

12

b.40

OB-Fold

80.4

68.4

13

b.42

β-Trefoil

89.0

100.0

14

b.47

Trypsin-like serine proteases

75.0

50.0

15

b.60

Lipocalins

90.5

100.0

16

c.1

TIM β/α-barrel

93.8

93.8

17

c.2

FAD/NAD(P)-binding domain

89.7

91.7

18

c.3

Flavodoxin-like

60.2

46.2

19

c.23

NAD(P)-binding Rossmann

90.2

85.2

20

c.37

P-loop containing NTH

79.5

50.0

21

c.47

Thioredoxin-fold

97.5

87.5

22

c.55

Ribonuclease H-like motif

75.3

58.3

23

c.69

α/β-Hydrolases

78.4

71.4

24

c.93

Periplasmic binding protein-like

92.0

100.0

25

d.15

β-Grasp (ubiquitin-like)

69.4

25.0

26

d.58

Ferredoxin-like

76.8

59.3

27

g.3

Knottins (small inhibitors, toxins, lectins)

88.2

96.3

Accuracy

  

85.3

76.2