Comparative analysis of co-expression networks reveals molecular changes during the cancer progression

Prostate cancer is a serious genetic disease known to be one of the most widespread cancers in men, yet the molecular changes that drive its progression are not fully understood. The availability of high-throughput gene expression data has led to the development of various computational methods for the identification of key processes involved. In this paper, we show that constructing stage-specific co-expression networks provides a powerful alternative strategy for understanding molecular changes that occur during prostate cancer. In our approach, we constructed independent networks from each cancerous stage using a derivative of current state-of-art reverse engineering approaches. We next highlighted crucial pathways and Gene Ontology (GO) involved in the prostate cancer. We showed that such perturbations in these networks, and the regulatory factors through which they operate, can be efficiently detected by analyzing each network individually and also in comparison with each other. Using this novel approach, our results led to the detection of 49 critical pathways and GOs related to prostate cancer, many of which were previously shown to be involved in this cancer. Correct inference of the processes and master regulators that mediate molecular changes during cancer progression is one of the major challenges in cancer genomics. In this paper, we used a network- based approach to this problem. Application of our approach to prostate cancer data has led to the re-establishment of previous knowledge about this cancer, as well as prediction of many other relevant processes and regulators.