The Effect of Edge Definition of Complex Networks on Protein Structure Identification
Figure 2
The flowchart of the amino acid interaction network together with community division method. PDB code 1BZ7, chain A. Each point in the figure represents an amino acid in the chain, which is also the vertex of the network. Ligatures between the vertices are the edge of the network, which illustrate the interaction between the amino acids. Here, the reddish purple and blue represent different domain regions in this chain based on the identity in SCOP. Firstly, an amino acid complex network was constructed with the vertex defined as , and the edge as - distance which was set at 7.5 Å, as shown in (a). Secondly, community division was based on edge betweenness, and the first edge with the highest edge betweenness score was removed, as shown in (b). Thirdly, more edges were removed based on the algorithm, and (c) shows that three edges were removed. Fourthly, the community division was finished when the correct number of edges was removed, as shown in (d); two different domains have been clearly separated, and five edges were removed for this protein. Finally, if the community division is taken continually, more communities will be found in the complex network. (e) shows the result of community division for chain A of protein 1BZ7 after removing 500 edges in this complex network, and many more communities illustrate the wrong results according to the identity in SCOP.
