We report on the integration of three methods that predict, on a proteome-wide scale, whether two proteins are likely to form a binary complex. Biological insights are derived from subnetworks, including the annotation of proteins of unknown function.
