Research On Protein Complexes And Essential Proteins Detection Based On Dynamic PPI Networks

With the rapid development of bioinformatics and high-throughput technology,it has brought a hot research topic that analysis comprehensively protein structure and function in the field of functional proteomics.The mining of protein complexes and the identification of essential protein can provide theoretical foundation to explains a law of process of life activities and reveal the lift`s mysteries.In addition,it has important significance to the drug research and the diagnosis of the disease.Currently,protein complexes and essential proteins detection algorithms are mature,but protein-protein interaction network has characters such as uncertainly,sale-free and small-world,as well as limitations of clustering algorithm,clustering precision and recall is low.In this paper,there are two contributions for identifying protein complexes by making use of intelligent optimization algorithm.On the one hand,combined with the characteristics and structure of PPI network,we construct new clustering model by simulating ant colony behavior.On the other hand,parameter of traditional clustering methods has been optimized by artificial bee colony algorithm.The precision of essential proteins prediction is improved by constructing temporal weighted network with dynamic and conserved proteins,integrating dual topology characteristics and biological properties.The main research of this paper is as follow:Since static PPI networks are difficult to truly reflect the dynamic character of cells,clustering precision is low,the convergence speed is slow in mining protein complex based on ant colony clustering,this paper proposes an ant colony clustering algorithm based on fuzzy granular and closeness degree to mine protein complexes in dynamic weighted PPI network,named FGCDACC.First,a comprehensive weight metric is designed based on the topological and biological characteristics of the PPI network.Second,this method constructs a series of dense and highly co-expressed complex core based on the module characteristic of the complexes,then it improves the picking and dropping operations,which based on fuzzy granular and closeness degree,to cluster the nodes in PPI networks,in order to speed up the clustering speed and improve clustering precision.Finally,this algorithm designs a weight's update strategy of positive feedback mechanism based on timing functional relevance theory and function transmission,which achieve the optimization transmission between different generations of ant colonies and networks at different times.Aiming at unreasonable global parameter settings,as well as to overcome thedisadvantage of slow convergence and being vulnerable to trap in local optima of artificial bee colony algorithm,this paper proposes a modified density-based clustering method based on improved artificial bee colony optimization,named IABC-DBSCAN.First,a truncation-championship selection mechanism and adaptive step and globally guided search strategies are proposed to enhance the diversity of the population and strengthen the ability that find the local solution of following bees.Then,the improved artificial bees colony is used to dynamic adjust DBSCAN `s optimal parameter,which is used as DBSCAN's input parameter.We run it on DIP data sets to detect protein functional modules,which verifies the proposed algorithm.At present,most of essential proteins identification algorithms are based on static PPI network,which neglect the inherent dynamics,conservative,as well as biological nature of proteins.In addition,those methods do not solve the problem of PPI data false positive and false negative.In allusion to the problems mentioned above,the temporal weighted network with dynamic and conserved proteins is constructed and a novel method called JTBC based on temporal weighted network is proposed to predict essential proteins.Firstly,according to gene expression data,dynamic information of dynamic proteins and conserved proteins is extracted to dynamically adjust static PPI network to construct temporal and conserved PPI network.Then we design node and edge cohesion coefficient,combine protein' domain information and Pearson correlation coefficient both with biological properties to weight all the interactions in temporal networks.Finally,we come up with co-expressed complex centrality based on the aggregation and co-expressed properties of essential proteins.Our algorithm is more reasonable to identify essential proteins from a global and local perspective,which comprehensively predicts essential proteins by integrating weight and complex information.