Abstract
BACKGROUND: MicroRNAs (miRNAs) are potential regulators that contribute to the pathogenesis of cancer. Microarray technologies have been widely used to characterize aberrant miRNA expression patterns in cancer. Nevertheless, the miRNAs expression signatures identified for a same cancer differs among laboratories due to the cancer heterogeneity. In addition, how the deregulated miRNAs coordinately contribute to the tumourigenic process of prostate cancer remains elusive. RESULTS: We evaluated five outlier detection algorithms that take into account the heterogeneity of cancer samples. ORT was selected as the best method and applied to four prostate cancer associated microRNA expression datasets. After microRNA target prediction and pathway enrichment mapping, 38 Gene Ontology terms, 16 KEGG pathways and 99 GeneGO pathways are found putative prostate cancer associated. Comparison with our previous studies, we identified two putative novel pathways important in prostate cancer. The two novel pathways are 1) ligand-independent activation of ESR1 and ESR2 and 2) membrane-bound ESR1: interaction with growth factors signalling. CONCLUSIONS: We proved that expression signatures of at the pathway level well address the cancer heterogeneity and are more consistent than at the miRNA/gene levels. Based on this observation, we identified putative novel microRNA regulatory pathways which will help us to elucidate the cooperative function of different microRNAs in prostate cancer.