Skip to main content

Table 3 Interface prediction accuracy, with heterocomplexes divided into functional classes

From: Conserved residue clusters at protein-protein interfaces and their use in binding site identification

Interface type (number)

Number (and percentage) of interfaces with Rank 1

 

Method 1

Method 2

Method 3

Homodimers (121)

65/121 (53.7)

[54/121 (45)]

58/121 (47.9)

51/121 (42.2)

Heterocomplexes (389)

189/389 (48.6)

[180/389 (46.3)]

196/389 (50.4)

187/389 (48.1)

   Enzyme-inhibitor (114)

77/114 (67.5)

80/114 (70.2)

75/114 (65.8)

   Antibody-antigen (41)

10/41 (24.4)

09/41 (22.0)

10/41 (24.4)

   Signaling complexes (78)

32/78 (41.0)

34/78 (43.6)

32/78 (41.0)

   Others (156)

70/156 (44.9)

73/156 (46.8)

70/156 (44.9)

  1. Details of the slightly different methods used to generate the surface patches are provided in Methods. Eq. 1 was used to calculate sequence entropy; however, in a few cases (values in square brackets) calculations were also repeated using Eq. 1a.