Identifying Consistent Control Paths And Its Application To The Identification Of Key Pathways In Cancer

With the rapid development of human society,complex systems from technology,biology,finance and society have sprung up.In order to meet the need of controlling complex systems,a large number of studies on network controllability have achieved fruitful results.However,most of them take nodes in the network as research objects,with few related to paths.Moreover,network controllability has been so far characterized only on isolated networks,while the vast majority of complex systems are formed by multilayer networks.On the other hand,with people's attention to human health,cancer pathway identification has gradually become a hot topic.Nevertheless,most methods for identifying cancer pathways are unable to analyze various cancers simultaneously and systematically,and thus cannot find common dysfunctional pathways across multiple cancers.Considering the relevant problems in existing studies,based on multilayer network models and network controllability,this paper proposes a graph pattern called consistent control paths.To identify consistent control paths in multilayer networks effectively,we propose an algorithm based on random sampling.The algorithm is tested on synthetic data to explore the influence of topological characteristics on consistent control paths,and the behavior of consistent control paths under random perturbations.By virtue of multi-omics data in TCGA,we apply the proposed algorithm to the identification of cancer key pathways.Compared with the cancer pathway found by most existing methods,the cancer key pathways identified in this paper can reflect the universal pattern of oncogenic signals diffusing in biological systems under different cancer environments.Through a lot of verification and analysis on synthetic data and real data,the algorithm designed in this paper can effectively discover the consistent control paths in multilayer networks.In addition,this method provides a new tool and platform for the study of controllability on multilayer networks.The results show that the consistent control paths are only related to the number of layers and the consistency of edges among different layers in a multilayer network.After analyzing the result of random perturbation,the removal of edges has greater influence than the reverse of edges has on consistent control paths.In the application to the identification of cancer key pathways,the identified cancer key pathways not only have remarkable biological significance,but also is helpful to reveal the common and specific oncogenic mechanisms during the development of different cancers.Furthermore,the cancer key pathways might play a guiding role in drug repositioning.Therefore,the method for identifying consistent control paths has offered a platform for analyzing different cancers systematically.