Studying On The Relationship Between Microbial Metabolic Networks And Resistance Based On Intelligent Computation

As the result of thermophilic bacteria adapt in high temperature environment, in recentyears their thermophilic enzymes are more and more applied in industrial production. Atpresent, the research of the heat stability influence factors of thermophilic bacteria havefocused on the analysis of genome and structure analysis of thermophilic enzymes. In order tolearn the resistant mechanism of thermophilic bacteria from the system perspective, deeplyunderstand the microbial resistance factor, and with the purpose of applications ofthermophilic bacteria in industrial. The whole metabolic network of the microbial as theresearch object, analyzed microbial metabolic network's topological structure properties, themetabolic network modularity and alignment, thereby studying the thermophilic bacteriaresistance influencing factors.The task of analysis metabolic network's topology and functions is to reconstruct thenetwork. We first got all metabolic reactions and other data involved in the metabolicnetworks from KEGG and DSMZ. We reconstructed two type networks. First is network ofinteracting pathways in which the nodes are metabolic pathways and the links are thepathway—pathway interactions. Second is metabolite graph, removed the current metabolites,with metabolites as nodes, corresponding to metabolic reaction as edges.Subsequently, using NetworkX to calculate 22 commonly used network topologyproperties of the sample biological metabolic networks. Using the factor analysis methodobtained 11 key factors of the metabolic networks on the microbial heat resistance. Then withthe 11 network characteristic attribute for the feature vector, respectively by genetic algorithm,particle swarm optimization algorithm and grid search algorithm to optimize the parametersof LS-SVM, further analysis the key factors of microbial resistance. Final averageinformation on clique size distribution etc. 11 network properties are the key factors ofmetabolic networks on the microbial heat resistance.Then, in order to further understanding the functional structure of microbial metabolicnetworks, respectively using bow-tie structure, hierarchical clustering,"component"definition an Girvan-Newman algorithm modularity on sample biological metabolismnetworks. Though the analysis found that metabolic networks of mesophilic bacteria are morecomplex than thermophilic, metabolic networks'internal modules of the thermophilic bacteriaare closer than mesophilic.Finally, in order to find the difference of metabolic network structure betweenthermophilic bacteria and mesophilic bacteria, we compared the microbial metabolic networks with random network models, and found that metabolic network of thermophilic bacteria issimilar with ER model and STICKY model, but metabolic network of mesophilic bacteria ismore similar to GEDGD models. We employed GRRAL and MI-GRRAL methods to alignthe metabolic networks of thermophilic and mesophilic bacteria, the results suggested thatnodes belonging to the network that effecting of microbial heat resistance are a small part ofthe non—hub nodes.The 11 selected metabolic network properties are the key factors of metabolic networkson the microbial heat resistance. The microbial metabolic networks modularity contributes todiscovery the different functional modules of thermophilic and mesophilic Network alignmentis the need for knowledge transfer; the result can further determine the key nodes of thenetwork of affect the thermophilic bacteria, providing theoretical reference to theexperimental method to determine the microbial resistance mechanism.