From: Protein complexes identification based on go attributed network embedding
Algorithm 1 Protein complex identification algorithm GANE |
---|
Input: Graph G = (V, E), GO property matrix O, vector representation dimension d, threshold value θ |
Output: A set of discovered protein complexes |
Description: |
Constructing a protein attribute affinity matrix S∈Rn×n |
Generating vector representation for each protein φ∈Rd |
Constructing a weighted adjacency matrix W |
Initializing Alternative_core, Seed_core, ComplexSet to be \( \varnothing \) |
Generating maximal cliques and put them into Alternative_core |
While \(\textit {Alternative\_core}\ne \varnothing \): |
DescendSort(Alternative_core) by density_score |
Alternative_core=Alternative_core−Clique1 |
Seed_core=Seed_core+Clique1 |
Pruning and updating remaining cliques in Alternative_core |
End while |
For core corei in Seed_core |
finding the set of its attachments Atti |
ComplexSet=ComplexSet+ corei∪Atti |
End for |
Return ComplexSet |