Research On The Detection Of Functional Modules In Protein Interaction Networks

The study of protein complexes and protein functional modules is an important way to further understand the mechanism and organization of biological activities.Research on the clustering methods and other computational methods used to analyze the information contained in protein-protein interaction(PPI)network is an effective way to explore the characteristics of protein functional modules.Based on the topological properties of PPI networks,this paper analyzes the problem of low detection efficiency in the overlap structure and noise of protein functional modules.The main contents of this paper are as follows:1)A module overlapping structure detection method in PPI using an improved link similarity-based Markov Clustering Algorithm(MLS)is proposed.One significant characteristic of PPI network is modularity,and the topological features of known functional modules are analyzed.It is concluded that there are overlapping structures between the modules.The algorithm can effectively detect the overlapping structure between functional modules in PPI networks.We test our method on several PPI network datasets of yeast proteins,and the clustering results were analyzed.It was found that the performance of the PPI algorithm was better than that of the MCL,LC,MCODE,Newman and other algorithms.Through analyzing the examples,it further shows the functional modules detected by the algorithm are more biologically significant.2)A protein functional module detection method based on topological features and gene expression data is proposed.By analyzing the topological structure of the known protein functional modules,it is found that there are noise data in them,and it can be seen that the data of protein interaction cannot fully reflect the relationship between proteins in the functional modules.Therefore,we introduce gene expression data and the topological features of PPI network to cluster analysis of PPI network.The experimental results of this algorithm are compared with those of other algorithms.It is concluded that the protein function module can be detected better by using the proportion ECTG algorithm.3)The comprehensive analysis system of PPI network clustering methods is designed and developed.The platform currently integrates eight clustering algorithms such as MLS,ClusterONE,ECTG and other clustering methods and seven evaluation methods such as Avg.F.The platform displays the comprehensive analysis results in a visual way,such as charts,which can quickly and effectively compare the performance of multiple algorithms.The research contents above are the main work of this paper,and the results can be used as an effective method to detect protein functional modules in PPI networks.