The Research On Algorithm Of Identifying The Functional Modules Based On The MicroRNA Regulatory Network

MicroRNAs(miRNAs)are a family of small non-coding RNAs that play an essential role in many biological processes by regulating the target genes(mRNAs),especially in the initiation and development of cancers,because the abonrmaloty of miRNAs function would lead to the initiation of the different complex diseases,and possibly even cancers.The research of miRNA-mRNA functional regulatory modules is one of the important ways to understand the regulatory mechanism of miRNAs,therefore,the identification of the miRNA-mRNA regulatory modules is an important breakthrough for understanding the pathogenesis of human complex diseases.However,most of the computational methods for constructing the regulatory networks were based the statistical correlations of miRNAs and the target genes,and these relationships indicated that miRNAs and genes have strongly correlated across samples and they could not represent the real regulatory relationships.In fact,it was proved that the relationship of miRNAs and their target genes might belong to the causal relationship.Moreover,some researcher have presented the network flooding model of regulatory network to reflect the regulatory between miRNAs and the target genes better.To reflect the regulatory between miRNAs and the target genes better,this paper present two different algorithms of identifying the miRNA-mRNA functional regulatory modules by integrating the causal relationship and the network flooding,the main work are as follow:(1)This paper proposes a novel approach called CALM(the causal regulatory modules)to identify the miRNA functional modules based on the causal interactions.The CALM algorithm first uses the IDA strategy to identify the causal miRNA-mRNA relationships by integrating the gene expression profiles and the information of the target bindling sites;then constructs the miRNA-miRNA synergistic network based on the relationship of miRNAs which can capture by computing the GO function similarity of their target genes;finally,detects the miRNA clusters and expands each miRNA cluster by greedy adding or discarding the target genes to form the miRNA-mRNA regulatory modules.The CALM algorithm can reduce the false positive of the miRNA-mRNA regulatory relationships identified by the statistical correlations.The experiment results show that the CALM algorithm achieves ideal overall performance in terms of the functional enrichment.(2)To impove the effectiveness of the algorithm in sparse network and reflect the regulatory between miRNAs and the target genes better,this paper proposes a novel approach called FLDM(the flood-module)to identify the miRNA-mRNA functional modules based on the network flooding.The approach first constructs the causal miRNA-mRNA regulatory network based on the previous work,and builds the weight network by computing the weighted of miRNA-mRNA with a new strategy;and then uses the theory of network flooding to translate the causal miRNA-mRNA regulatory network to the flood network of miRNA-mRNA,forms the candidate sub-network of miRNA-mRNA by minimizing the flood network;finally,merges or discards the sub-network to get the miRNA-mRNA functional regulatory modules.The experiment results also show the effectiveness of FLDM in sparse network and the overall performance in terms of the functional enrichment.