MicroRNAs (miRNAs) are often dysregulated in ovarian cancer. These short non-coding RNA sequences regulate gene expression by repressing specific, targeted mRNAs. We investigated the effects of miRNAs, found to be upregulated in ovarian cancer, on thousands of genes, and found that certain miRNAs appeared to be critical to the transformation of ovarian epithelial cells from normal to cancer. In addition, we found evidence that the deactivation of powerful transcriptional repressors often results in the unexpected upregulation of predicted miRNA targets. We also demonstrated, computationally and empirically, that even a single nucleotide mutation in the miRNA"seed region" (that portion of a miRNA that specifies which mRNAs will be targeted) can result in the complete disablement of that miRNA's repressive ability against its putative targets. We built a phenomenological network model by comparing changes in expression between millions of pairs of genes in normal and cancerous ovarian epithelial cells. We found that the number of highly correlated gene pairs decreased dramatically between normal and cancer tissues. This suggests either a loss of control in cancer, the existence of sub-populations within the tumors, or the existence of multiple tumor subtypes. We noticed that most pairs of genes which were highly correlated in normal tissue remained correlated in cancer. Gene pairs which did not maintain their correlation in cancer, we termed inconsistent. Inconsistent gene pairs shared some important characteristics: 1) one of the genes was generally upregulated; 2) both genes often were enriched for the cell cycle and/or apoptotic processes; and 3) at least one member of the gene pair had been previously implicated in ovarian and/or other cancers. The complexity of these highly interconnected network modules suggests that there are no master regulators of cancer. Rather, it implies that there exists an orchestra of dysregulated genes which interact in small ensembles, disrupting critical biological processes, motivating proliferation and metastasis.
