Research On Key Genes Identification Methods Based On Multilayer Network

Identifying key genes based on their topological characteristics in biological networks are essential for further understanding of the pathogenic mechanism of organisms.For the problem of identifying key genes through network centricity in incomplete interaction networks such as the Human network.This paper applies the random walk algorithm to the identification of key genes from the perspective of multi-layer biological network.The peculiar essential genes with small degree can be revealed in Yeast-Human interconnected networks.The main work of this paper is as follows:(1)The relevant theories of multi-layer network are further studied.Based on the expansion of single layer network,multi-layer network can be used to model multiple interrelated systems,effectively integrate different interaction networks,and can quantify the role of nodes from multiple directions systematically.In view of the existence of homologous gene pairs between biological interaction networks,a multi-layer biological network,which can be called biological network integration,has been constructed to effectively solve the interference of data noise,deviation,inconsistency and different dimensions in the experiment.(2)The key genes identification methods on complex network are researched profoundly.By analyzing the principle and advantages and disadvantages of the identification methods of key nodes,because random walk model can quantify the long-term influence of nodes,and the topology characteristics of key genes are revealed by using random walk model on the constructed multi-layer biological network.The Random Walk Occupation(RWO)of nodes after random walk is used as the centrality index of multilayer network to identify the key genes.The recognition method is designed to identify the key genes with small degree via orthologous key genes with high degree.The multi-layer network constructed in this paper has effectively integrated two different biological interaction networks with different topology structures.On this basis,the method of combining with random walk model has been used to achieve better experimental results on key genes recognition.By objective comparison and verification,the superiority of the method is fully illustrated,and it is also an extension and application of graph theory,algorithm,and technology.