Figure 1

Flowchart of the combinatorial approach to determine the transcriptional and posttranscriptional regulatory motifs based on gene expression profiles. Firstly, we conducted a genome-wide screening to detect potential miRNA target motifs in Arabidopsis based on an inhomogeneous HMM and cross-species conservation and minimum binding energy of miRNA/mRNA duplex were used as additional filters to reduce the rate of false positives. Secondly, genes in the cop1 mutant time course microarray dataset were clustered into 12 expression patterns and overrepresented sequence elements in the upstream of the genes belonged to the same cluster were detected using AlignACE. Thirdly, Bayesian network strategy was applied to selecting these motifs in both upstream sequences and transcripts that were most related to the gene expression patterns. Lastly, we measured the degree to which gene expression could be determined merely by these adopted regulatory motifs.