Regulation and functional roles revealed by clustering of microarray expression data ofEscherichia coligenes

An enormous amount of gene expression data is currently available online in repositories for several organisms. Microarray data can be used to identify co-expressed genes that may be involved in the same biological process. Therefore, the analysis and interpretation of this information could help organize and understand the knowledge it contains, representing a major challenge in the post-genomic era. Here, we grouped genes ofEscherichia coliK-12 using expression data to infer meaningful transcriptional regulatory information. Our method assumes that co-expressed genes reflect functional units, as evidenced by their genetic structure, including gene arrangement, regulation, and participation in defined biological processes. These functionally linked clusters were validated with curated transcriptional regulatory information from RegulonDB. From 907 growth conditions, 420 clusters were formed involving 1674 genes. Clusters contained from 2 to 64 genes. We found that co-expressed genes participate in related metabolic pathways and share similar types of regulation (through transcription factors,σ-factors, allosteric regulation, or micro-RNA regulation). This study is helpful for identifying novel transcriptional regulatory interactions.