Research And Visualization Of Annotation-based Modular Recognition Method In Biological Networks

With the development of high-throughput biotechnology and increase size of biological data,biological networks become an important data structure to represent relation data between biological systems and biological entities.In the study of biological networks,researchers found that the modular structure of networks is often of practical biological significance,therefore,the modular recognition has become an important method in biological network analysis.The mainstream algorithms mainly focus on the topological structure information.However,for biological networks and other real-world networks,nodes also have abundant label information.It is a new research direction to identify network modularity by combining node annotation and network topology information.Therefore,this topic aims to study the current classical and advanced annotation-based network modular identification methods,and selects six algorithms to compare their effects and performance.At the same time,the application in the biological network labeled GO annotation is implemented.Due to the particularity of the hierarchical structure of the GO annotation,the repetition rate of the label information carried by the network node is low,the length is not fixed,and the amount of attributes are large.Aiming at these of problems,we proposed an annotation redefinition method based on GO topology partitioning.In this method,we compared the simple topology-based modular identification method and the label-based modular recognition method in the biological network.In order to comprehend module partitioning data more easily,we developed an online visualization platform for modular identification of biological networks,which provides functions including data input,label-based modular identification and visualization display,so that provides users with great convenience.