Fig. 2.

a–c Interaction site recognition: visualization of 3D structures of representative heterocomplexes. The target protein in each complex is shown in strands, with residues of interest shown in space fill and color-coded as follows: gray, true positives (interface residues identified as such by the classifier); white, false negatives (interface residues missed by the classifier); black, false positives (residues incorrectly classified as interface). The interaction partner is shown in gray wireframe. a FabN10 in the 1nsn complex. b α-chymotrypsin in the 1acb complex. c Elastase in the 1fle complex. Structure diagrams were generated using RasMol (http://www.openrasmol.org/)