Skip to main content

Table 5 Comparison of the present work with previously published algorithms, Benchmark 2

From: Identifying structural domains of proteins using clustering

Algorithm1

Time2

1-domain

2-domain

3-domain

4-domain

Overall3

SS Algorithm

5s

86% (0.75)

64% (0.47)

60% (0.46)

0% (−0.03)

69%

CA Algorithm

26s

92% (0.82)

78% (0.69)

76% (0.69)

25% (0.32)

80%

DDomain

497s

94% (0.78)

75% (0.68)

48% (0.56)

25% (0.16)

75%

DomainParser2

398s

98% (0.86)

75% (0.71)

64% (0.60)

50% (0.39)

79%

PDP

99s

92% (0.93)

84% (0.82)

68% (0.69)

75% (0.55)

83%

Random (see text)

1s

71 ± 3% (<0)

15 ± 7% (<0)

1 ± 2% (<0)

0% (<0)

31 ± 3%

  1. Present work compared to DDomain[12] (using AUTHORS parameters), DomainParser2[13] and PDP[11] on the Benchmark 2 data set. Matthews correlation coefficient is given in parentheses. 1CA refers to the α-carbon based algorithm and SS the secondary structure based one. 2Time taken for the actual execution of the binaries or detection algorithms over the full dataset on a single 2.93 GHz i7 CPU. 3Total, regardless of number of domains.