A Selection-based MCL Clustering Algorithm for Motif Discovery

As motif discovery plays an important role in the understanding of the relationship of gene regulation, this paper puts forward a selection-based MCL clustering refinement algorithm (SMCLR) aiming at solving the planted (l, d) motif search (PMS) problem. Firstly, we divide the DNA dataset into different subsets through selection of reference sequence and screen parts of eligible subsets by setting thresholds under selection project. Then MCL clustering algorithm is used for refinement. The experiment resulted on simulation data shows that SMCLR algorithm has higher prediction accuracy in a reasonable time than these existing motif discovery algorithms like Project, MEME, MCL-WMR and VINE. Moreover, the experiment resulted on real biological data demonstrates the effectiveness of SMCLR algorithm.