Protein-Protein interactions (PPIs) are an important area of research because most cellular processes are carried out by aggregates of interacting proteins. Identification of these aggregate complexes can provide insight into how the proteome is organized into functional units \highlight{(A.-C. Gavin and et al., Nature, 415(6868):141-147, 2002) }. Proteins can then be leveraged for such uses as therapeutic drug design where PPIs are often mediated or blocked by the drug. This thesis is concerned with predicting binding sites that mediate a PPI by using re-occurring polypeptide sequences on a proteome scale, without the use of 3D structure data or gold-standard interaction site mediators. In this thesis, an algorithm has been developed to identify potential interaction sites, as well as a metric for evaluating the predictions. \highlight{This research shows } that interaction sites can be predicted from using re-occurring polypeptide sequences on a proteome scale for both human and yeast organisms.