Construction And Analysis Of Disease Specific Regulatory Networks

With the development of bioinformatics and genomics research,there are more and more studies on the impact of regulation and gene expression in life activity.The development of biotechnology like sequencing technology provides a large number of data for bioinformatics research.As main regulators in gene expression,transcription factor(TF)and micro RNA(mi RNA)play important roles in life process,which promote or inhibit gene expression separately in transcription and translation level.The differential expression of most genes produces proteins,which act as structure component and regulates metabolic activity as well as related to the cell survival and cell function.Abnormal expression of coding gene and non-coding gene leads to the initiation and progression of complex disease such as cancer,so that influence health and life of human.The regulation of gene expression is critical to the prevention and therapy of disease.Therefore,study the TF-mi RNA regulatory network with gene expression and protein-protein interaction(PPIs)is important to the study of disease pathomechanism.Combined with multiple types of data,this study proposed the process of construction and analysis disease specific regulatory network in molecular level.At first,regulatory relationship between TF,mi RNA and their targets,as well as PPIs were used to construct basic regulatory network.Then the expression data of specific disease in normal samples and tumor samples was used separately to calculate correlation between the nodes which have interactions in basic regulatory network,and the significant interactions were chosen to construct normal regulatory network and tumor regulatory network.Based on these two networks,a core nodes identification method was proposed to identify core nodes in the networks with consideration of the changes of nodes in architecture and topology.We also combined normal and tumor regulatory network and got disease specific regulatory network,and analyzed the relation between network and disease by pathway,function and disease.At last,based on the characteristic of disease specific network,we improved the K nearest neighborhood classifier and proposed the nearest neighborhood method to predict disease candidate nodes,which can be used to valid the relationship between disease and molecular in priority.The expression data of lung adenocarcinoma(LUAD)was used to construct and analyze LUAD specific regulatory network.All the networks in this study are scale-free network,which is in consist with real biological network.The analysis demonstrated that the method proposed in this paper could identify core nodes effectively,which are closely related to human lung disease.The candidate molecular predicted by the nearest neighborhood method has the higher precision than candidates predicted by random walk with restart.The enrichment analysis of nodes in disease specific regulatory network and disease candidate nodes showed close relationship with lung disease.In one word,The construction and analysis process processed to construct disease specific regulatory network can identify core nodes in the network and predict disease candidate molecular,which can be used in other complex disease and contribute to understanding the complex disease pathomechanism.