Complex Network Analysis Reveals Biomarkers In Lung Cancer

Non-coding RNAs play an indispensable role in numerous biological processes ranging from organismic development to tumor progression.In oncology,these RNAs constitute a fundamental component in the competing endogenous RNA(CeRNA)hypothesis that provides the basis for probing into the mechanisms and evolution of various cancers through the complex networks.Previous work in this area focused on static CeRNA networks.During the development of cancers,the underlying CeRNA network would hardly remain unchanged,so we investigated the dynamical changes of the CeRNA network.Taking the lung adenocarcinoma(LUAD)as a prototype case of study,we construct and analyze the CeRNA networks for four distinct sequential stages of LUAD evolution based on multi-omics data of long non-coding RNAs(lncRNAs),microRNAs and mRNAs.For any given stage,a quantitative approach to reconstructing the CeRNA network consists of analyzing differentially expressed RNAs,matching microRNA targets by the principle of base complementary pairing,and selecting the negative correlation of RNA expression by invoking the CeRNA hypothesis.We found that the networks comprise microRNAs and target mRNAs(or lncRNAs),and possess a two-level bipartite structure.Comparing the components of microRNAs among the networks at the four stages of LUAD,we uncover two types of subnetworks as a natural partition of the network: the common competing endogenous networks(CCENs)composed of an invariant set of microRNA over all the stages,and the unique competing endogenous networks(UCENs)that are characteristically distinct for different stages.A systematic enrichment analysis of Gene Ontology(GO)and Kyoto Encyclopedia of Genes and Genomes(KEGG)pathways of the mRNAs in the CCENs are carried out.As LUAD progresses from stage I to IV,the mRNAs in the CCENs exhibits a strong association with the cancer development.With respect to the mRNAs selected directly from CCENs or UCENs,the microRNAlinked mRNAs from the UCENs have a higher enrichment efficiency.The principal component analysis technique studies the LUAD sample data after differential expression analysis,and combines the Cox proportional hazard regression model to analyze the effects of each principal component and its eigenvectors on the survival of LUAD patients.Through hierarchical clustering of eigenvectors of PCA,we found that some microRNAs reported in the work of CeRNA network can significantly affect the survival of LUAD samples.Through Kaplan-Meier survival curve analysis,a key finding is six microRNAs from that CCENs that impact the survival of LUAD patients at all stages,and four microRNAs that affect the survival at a specific stage of LUAD.These ten microRNAs can then serve as potential biomarkers and prognostic tools in the diagnosis and monitoring of LUAD.