Searching Transcriptional Modules Using Evolutionary Algorithms

The mechanism of gene regulation has been studied intensely for decades. It is important to identify synergistic transcriptional motifs. Its search space is so large that an efficient computational method is required. In this paper, we present the method that can search automatically both transcriptional motif list and gene expression profiles for synergistic motif combinations. It uses evolutionary algorithms to find an optimal solution for the problems which have the huge search space. Our approach includes the additional evolutionary operator performing local search to improve searching ability. Our method was applied to four Saccharomyces cerevisiae gene expression datasets. The result shows that genes containing synergistic motif combination from our optimization technique are highly correlated than those from k-means clustering. In cell cycle as well as other expression datasets, our results generally coincide with the previous experimental results.