Skip to main content

Table 4 The Q3 accuracy comparison of SymPred using dictionaries compiled from different percentages of the template proteins.

From: Improving protein secondary structure prediction based on short subsequences with local structure similarity

Percentage of template pool

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Number of template proteins

830

1660

2490

3320

4150

4980

5809

6638

7467

8297

Q3 on DsspNr-25

70.8

73.6

75.0

76.3

77.3

78.1

78.7

79.3

79.8

80.5

Improvement

-

+2.8

+1.4

+1.3

+1.0

+0.8

+0.6

+0.6

+0.5

+0.7

  1. The performance improves as the number of template proteins increases. SymPred’s performance improves between 0.5% and 2.8% each time the number of template proteins is increased by 10%.