The most over represented processes in Cluster 4 genes were granulocyte mediated immunity, NF B and cytokine and chemokine signaling. We further analyzed Clusters 2, 3 and 4 using network analysis to discover transcriptional regulatory modules that could potentially be http://www.selleckchem.com/products/wortmannin.html responsible for coordinate reg ulation of these three clusters. We observed that in Clusters 2, 3 and 4 there were common hubs of transcriptional control. p53 and NF B proteins were potential transcription factors of genes in all three clus ters, which had similar overall profiles. KDM5B JARID1B was once again identified as a poten tial upstream regulator of genes in both Clusters 2 and 4. In Cluster 4, the genes potentially regulated by KDM5B were the same as those in FBPA Cluster 3 after irradiation, and in bystanders KDM5B was also shown to be upstream of the GADD45A and SAT1 genes in Cluster 2.

It was interesting that the metallothionein gene expression response in bystanders was similar to that in irradiated cells, suggesting that irradiated cells may be communicating a signal that induces epigenetic changes in both populations. Protein analysis on KDM5B, HDAC1 and HDAC2 levels showed that these histone modifiers are lowered in bystanders at 1 hour after treatment as in the directly irradiated cells. This suggests that the bystander metallothio nein gene response maybe regulated similarly as in irradiated cells. Biological and statistical evaluation of clustering results In terms of the clustering methodologies used here, the most surprising result was the high degree of biological information found using the FBPA clustering versus STEM clustering across both cell treatments, despite roughly equivalent computational evaluations.

We observed similar enrichment results for other STEM clusterings of the data with various parameters. Although there were some common processes between FBPA clusters, the gene ontology enrichment showed clear delineation of biological information. Related biological functions were focused in specific clusters, suggesting that features used in FBPA captured relevant biological details of the gene expression response curves. In radiation gene response, three out of four clusters gave distinct functional groups, a cell signaling cluster, a cell cycle cell death cluster and a cell mediated immunity cluster. Network analysis clearly Anacetrapib revealed the differences in individual players and suggested novel regulatory mechanisms for the coordi nate responses. By contrast, STEM resulted in only one cluster with biologically significant functions for both treatment conditions, irradiated Cluster 3, and bystander Cluster 1, which encompass processes from signal transduction modules, cell specific immunity, cell death and cell proliferation responses.