4 Application of Protein–Protein Docking to Water-Soluble, Membrane-Anchoring,
and Integral Membrane Proteins
Most protein–protein docking software was tailored for water-
soluble proteins (Fig.1). There are a number of attempts to evalu-
ate protein–protein docking software for this purpose. In particu-
lar, the protein–protein docking benchmark 4.0 by the Weng group
[37] was used to evaluate the docking performance of available
software. The benchmark has a total of 176 diverse targets includ-
ing 52 enzyme–inhibitor, 25 antibody–antigen, and 99 cases with
other function. Huang et al. [5] performed a comprehensive assess-
ment of the 18 docking/scoring protocols of 14 global dockingFig. 1Example of protein–protein interaction interfaces considered in particular classes of proteins. (a)
Transmembrane protein dimerization involves only transmembrane domains, which restricts interface search
to these regions only (yellow), which can facilitate docking as long as the software allows for appropriate
settings. On the other hand, interactions of transmembrane proteins with soluble proteins are likely to involve
the solvent-exposed regions only (blue; on the example of a GPCR, discussed farther in the main text, PDB ID:
4dkl). (b) Also dimerization of membrane-anchored proteins is to some extent facilitated by supposed
restriction of possible dimerization interface localization (yellow; on example of K-Ras protein investigated
by Prakash et al., discussed in the text, PDB ID: 4lv6). (c) In turn, soluble proteins are surrounded by solvent,
and a binding partner can approach from any side (the entire surface can potentially bind proteins, which is
indicated by yellow surface; on example of viral capsid protein investigated by Antal et al., which is discussed
in the text, PDB ID: 1za7)
290 Agnieszka A. Kaczor et al.