Mining Regulatory Elements in Non-coding Regions of Arabidopsis thaliana

Analysis of regulatory elements (DNA motifs) in non-coding regions is considered as one crucial step to understand the regulation mechanisms of genes with similar expression patterns. With the help of accumulated gene expression data and complete genome sequences, computational approaches have been developed in the past decade to accelerate the mining task. In previous studies, we proposed a DNA motif discovery framework, named as MODEC, which incorporated the evolutionary computation (EC) searching algorithm with data filtering techniques to favor the algorithm performance. With the attempt on exploring real-world motif mining problems, we apply both MODEC and a famous discovery algorithm MEME to predict regulatory elements in different non-coding regions of co-expressed genes from the model plant Arabidopsis thaliana. Results from both MODEC and MEME show that the targeted motif patterns can be found in the expected non-coding regions of the co-expressed gene groups. As the preliminary step of this work, we investigate whether different motif patterns can be detected in the specified non-coding regions of co-expressed genes with different functional categories. The similar prediction results from MODEC and MEME demonstrate the potential of MODEC in the field of practical motif discovery.