"iRefR" allows the user to load any version of the consolidated protein interaction database "iRefIndex" and perform tasks such as: selecting databases, pmids, experimental methods, searching for specific proteins, separate binary interactions from complexes and polymers, generate complexes according to an algorithm that looks after possible binary-represented complexes, make general database statistics and create network graphs, among others.
The paper introducing iRefR (from 2011): http://bmcbioinformatics.biomedcentral.com/articles/10.1186/1471-2105-12-455