Inference of new biological knowledge, e.g., prediction of protein function,from protein-protein interaction (PPI) networks has received attention in thepost-genomic era. A popular strategy has been to cluster the network intofunctionally coherent groups of proteins and predict protein function from theclusters. Traditionally, network research has focused on clustering of nodes.However, why favor nodes over edges, when clustering of edges may be preferred?For example, nodes belong to multiple functional groups, but clustering ofnodes typically cannot capture the group overlap, while clustering of edgescan. Clustering of adjacent edges that share many neighbors was proposedrecently, outperforming different node clustering methods. However, since somebiological processes can have characteristic "signatures" throughout thenetwork, not just locally, it may be of interest to consider edges that are notnecessarily adjacent. Hence, we design a sensitive measure of the "topologicalsimilarity" of edges that can deal with edges that are not necessarilyadjacent. We cluster edges that are similar according to our measure indifferent baker's yeast PPI networks, outperforming existing node and edgeclustering approaches.
展开▼
