Table 5.

Protein complex detection algorithm.

Input : an unweighted network, a weighted network built via GO annotation and a training set Complex detection process: Step 1: construct the feature vector space for the complexes in the training set from the unweighted and weighted PIN networks and train the Regression model Step 2: find maximal cliques in the PIN by the Cliques algorithm    -rank the clique set C={C1, C2, ..., Cn} in descending order of the scores given by the Regression model    -for each clique Ci, check all the cliques (denoted as Cj) with lower scores, if Ci∩Cj > threshold, then remove Cj.    -output: the updated clique set Step 3: grow the cliques    -for each clique Ci, the set of its neighbors is denoted as N(Ci), do update operation as follows:       -check all the nodes in N(Ci)       -select vi∈N(Ci), which makes vi∪Ci achieve higher score given by the Regression model       -update Ci= vi∪Ci, N(Ci) = N(Ci) - vi          -repeat the update operation until there is no node vj in N(Ci) that leads to score(vj∪Ci) > score(Ci)       -output: the candidate complex set C = {C1, C2, ..., Cn} Step 4: filter the candidate complexes    -rank the candidate complexes in descending order of the score given by the Regression model    -for each candidate complex Ci, check all the candidates Cj with lower scores          -if overlap (Ci, Cj) > merg_thred             if score(Ci∪Cj) > score(Ci) do merge operation: update Ci = Ci∪Cj             else do remove operation: remove Cj from the candidate set output: the predicted complex set