Localizing potentially active post-transcriptional regulations in the Ewing's sarcoma gene regulatory network
1 Systems Biology Group, Life Science Research Unit, University of Luxembourg,162A Avenue de la Faiencerie, Luxembourg, L-1511, Luxembourg
2 Work done at: IRMAR, Université de Rennes 1, Rennes, France
3 UMR 6625, CNRS, Rennes, France
4 Genetics and Biology of Cancers, Institut Curie, Paris, France
5 Unité 830, INSERM, Paris, France
6 Service bioinformatique, Institut Curie, Paris, France
7 Unité 900, INSERM, Paris, France
8 Service bioinformatique, Mines ParisTech, Fontainebleau, France
9 DIMNP, Université de Montpellier 2, Montpellier, France
10 UMR 5235, CNRS, Montpellier, France
11 Symbiose project team, INRIA, Rennes, France
12 UMR 6074, CNRS, Rennes, France
13 IRISA, Université de Rennes 1, Rennes, France
BMC Systems Biology 2010, 4:146 doi:10.1186/1752-0509-4-146Published: 2 November 2010
A wide range of techniques is now available for analyzing regulatory networks. Nonetheless, most of these techniques fail to interpret large-scale transcriptional data at the post-translational level.
We address the question of using large-scale transcriptomic observation of a system perturbation to analyze a regulatory network which contained several types of interactions - transcriptional and post-translational. Our method consisted of post-processing the outputs of an open-source tool named BioQuali - an automatic constraint-based analysis mimicking biologist's local reasoning on a large scale. The post-processing relied on differences in the behavior of the transcriptional and post-translational levels in the network. As a case study, we analyzed a network representation of the genes and proteins controlled by an oncogene in the context of Ewing's sarcoma. The analysis allowed us to pinpoint active interactions specific to this cancer. We also identified the parts of the network which were incomplete and should be submitted for further investigation.
The proposed approach is effective for the qualitative analysis of cancer networks. It allows the integrative use of experimental data of various types in order to identify the specific information that should be considered a priority in the initial - and possibly very large - experimental dataset. Iteratively, new dataset can be introduced into the analysis to improve the network representation and make it more specific.