Research On Prediction Model Of Disease Related-factor Based On The Bipartite Network

After the completion of human genome sequence map,the new sequencing technology has entered into the stage of rapid development and produced a large amount of biological and medical data.Facing the vast of data,it is of great meaningfulness to explore disease pathogenesis,diagnosis,prognosis,and treatment from different aspects including biomolecules,environmental factors and microbes.Using bioinformatics to predict disease-associated factors is currently one of the hotspots,which attracts researchers’ much attention and in-depth research.In recent years,the prediction models of network based disease related factors have attracted much attention and achieved good prediction results.The basic ideas of this kind of prediction model are as follows:first,the network model is constructed based on the known relationships.Secondly,the association between nodes is established by using part or all of the information of the network,and then the potential relationship is predicted.Its advantage is to use the relevant biological data and the topology characteristics of the network to predict disease associated factors.which is of great significance and worth continuing further research and analysis.In this thesis,we illustrated a simple introduction for predicting disease-associated factors,then we focus on the problems of predicting potential disease-LncRNA and disease-microbes relationships.And the detailed contents of the study are described as follows:(1)Prediction of LncRNA-disease association:first,we constructed a bipartite network by collecting the information of disease-related LncRNA.Secondly,the assumption,two nodes are similar if they have common neighbors or are connected to similar nodes,was utilized to calculate the correlations among diseases nodes and that of LncRNA nodes.Finally,we constructed a model for predicting the potential association between disease and LncRNA by combining the calculated results with the information of disease-related LncRNA.The experiments showed that the new prediction model is effective.(2)Prediction of Microbe-disease association:first,we transformed the information of known disease-associated microbes and genes into two bipartite networks.Secondly,the similarities among diseases and that of microbes were computed based on the newly constructed bipartite networks and the above assumption.Finally,a computing model for predicting potential associations between disease and microbe was designed by integrating multiple data sources.The experiments showed that the new prediction model is effective.(3)In the last chapter of this thesis,we analyzed the connection and difference between the above two forecasting models,made a conclusion for this paper and presented research prospects.