Research On Orientation Algorithm Of Protein-Protein Interaction Network

The biological signal transduction in the protein-protein interaction(PPI)network occupies a very important position in life activities,and it affects most of the biological activities of organisms.The study of the directionality of signal flow between proteins is helpful to understand and predict the response of cells to various signals,the function of genes and proteins in the cell,and the mechanism of internal environmental homeostatic.In recent years,with the widespread application of bioinformatics,the number and quality of PPIs have been significantly improved,but some PPIs are lacked information on the directionality of signal transduction,especially for complex human PPIs.In order to distinguish the oriented PPI from the unoriented PPI,and predict the directionality of signal flow between proteins,this research proposed an improved network propagation method,combined protein semantic similarity and overlapping clustering algorithm,for orientating PPI network.The specific research work is as follows:Firstly,this article made an unoriented VEGF signal transduction network,combined with the network propagation algorithm and protein semantic similarity,to study the performance of the network propagation algorithm in orienting the signal flow of PPI and distinguish the oriented PPI from the unoriented PPI.The results show that the network propagation algorithm has defect in distinguishing between oriented and unoriented PPI.Secondly,in view of the defect of the algorithm,we combined the clustering algorithm and network propagation algorithm for orienting the human integrin-mediated PPI network.As a result,484 unoriented PPIs(form protein complexes)and 682 oriented PPIs(interactions involving signal transmission)were obtained;For evaluating the effectiveness of the modified method,we compared with the original method and obtained better results;We extracted some signal transduction pathways from the directed integrin-mediated PPI network by using breadth-first search algorithm,and analyzed the extracted directional pathways.Finally,this article expanded the human integrin-mediated PPI network,which consists of473 proteins.In this network,1664 pairs of unoriented PPIs and 2556 pairs of oriented PPIs were obtained by the orientation algorithm.In order to evaluate the results of this network,we used GO function enrichment analysis to evaluate unoriented PPIs;ROC curve was used to evaluate the signal flow directionality of oriented PPIs,and the results were compared with that of another oriented method based on the shortest path.The results show the reliability of this research method.In summary,this research improves the network propagation algorithm in order to orient the human protein-protein interaction network.This algorithm distinguishes the unoriented PPI and the oriented PPI in the network,directs the signal flow of oriented PPI.Their directional information not only provide a basis for mining new signal transduction pathways,but also help promote the research of genome-related diseases,assist the development of drugs,and study unknown regulators of signal transduction pathways.