Detecting pan-cancer conserved microRNA modules from microRNA expression profiles across multiple cancers.

zhaowen liu,junying zhang,xiguo yuan,baobao liu,yajun liu,aimin li,yuanyuan zhang,xiaohan sun,shouheng tuo

DOI: https://doi.org/10.1039/c5mb00257e

2015-01-01

Molecular BioSystems

Abstract:MicroRNAs (miRNAs) play an indispensable role in cancer initiation and progression. Different cancers have some common hallmarks in general. Analyzing miRNAs that consistently contribute to different cancers can help us to discover the relationship between miRNAs and traits shared by cancers. Most previous works focus on analyzing single miRNA. However, dysregulation of a single miRNA is generally not sufficient to contribute to complex cancer processes. In this study, we put emphasis on analyzing cooperation of miRNAs across cancers. We assume that miRNAs can cooperatively regulate oncogenic pathways and contribute to cancer hallmarks. Such a cooperation is modeled by a miRNA module referred to as a pan-cancer conserved miRNA module. The module consists of miRNAs which simultaneously regulate cancers and are significantly intra-correlated. A novel computational workflow for the module discovery is presented. Multiple modules are discovered from miRNA expression profiles using the method. The function of top two ranked modules are analyzed using the mRNAs which correlate to all the miRNAs in a module across cancers, inferring that the two modules function in regulating the cell cycle which relates to cancer hallmarks as self sufficiency in growth signals and insensitivity to antigrowth signals. Additionally, two novel miRNAs mir-590 and mir-629 are found to cooperate with well-known onco-miRNAs in the modules to contribute to cancers. We also found that PTEN, which is a well known tumorsuppressor that regulates the cell cycle, is a common target of miRNAs in the top-one module and cooperative control of PTEN can be a reason for the miRNAs- cooperation. We believe that analyzing the cooperative mechanism of the miRNAs in modules rather than focusing on only single miRNAs may help us know more about the complicated relationship between miRNAs and cancers and develop more effective treatment strategies for cancers.