Evolutionary Bioinformatics

Network motifs are overly represented as topological patterns that occur more often in a given network than in random networks, and take on some certain functions in practical biological applications. Existing methods of detecting network motifs have focused on computational efficiency. However, detecting network motifs also presents huge challenges in computational and spatial complexity. In this paper, we provide a new approach for mining network motifs. First, all sub-graphs can be enumerated by adding edges and nodes progressively, using the backtracking method based on the associated matrix. Then, the associated matrix is standardized and the isomorphism sub-graphs are marked uniquely in combination with symmetric ternary, which can simulate the elements (–1,0,1) in the associated matrix. Taking advantage of the combination of the associated matrix and the backtracking method, our method reduces the complexity of enumerating sub-graphs, providing a more efficient solution for motif mining. From the results obtained, our method has shown higher speed and more extensive applicability than other similar methods.