Structural Controllability Analysis Of Static And Dynamic Transcriptional Regulatory Networks For Saccharomyces Cerevisiae

The controllability of complex network is the frontier of complexity science and one of the ultimate goal of complex networks.A network is controllable,if given an input,the system can reach desired state within limited time from any initial states.Although great progress on the controllability of complex networks has been made in recent years,few works have been done on the controllability of biological networks,especially for dynamic biological networks.The structural controllability of six transcriptional regulatory net-works?TRNs?for Saccharomyces cerevisiae?S.cerevisiae?were investigated in this paper,which included one static network and five dynamic ones.The five dynamic networks con-tain two endogenous and three exogenous ones.This thesis mainly obtained the following conclusions:Firstly,the static and dynamic TRNs of S.cerevisiae were constructed using the ex-isting literature.The theory of structural controllability allows us to classify nodes into five categories according to their control roles:driver nodes,critical driver nodes,in-dispensable nodes,dispensable nodes and neutral nodes.And the minimum number of driver nodes?MDS?used to analyze the structural controllability of static and dynamic TRNs were determined by the maximum matching algorithm.The results reveal that biological networks are rather difficult to be fully controlled,and it is necessary to control more than 74%of nodes in the network.However,the dynamic networks are relatively easier to be controlled than the static one,and one needs relatively more external inputs to control the exogenous networks than the endogenous ones.Subsequently,the compar-ison with the randomized networks and the statistical tests illustrate the validity of the conclusions.Secondly,the relationships between the feed-forward loop?FFL?,four-node single-input?SIM?and multi-input modules?MIM?and the controllability of the networks were considered.It was found that nodes in each network that consisting of the SIM and MIM tend to be critical driver nodes,and those consisting of the FFL in dynamic networks are prone to be critical drivers in comparison with those in the static network.Finally,the relationship between structural controllability and gene functions were analyzed by GO enrichment analysis and_?80?enrichment analysis.Enrichment analysis reveals that driver nodes but not critical driver nodes contribute to functional differences between the static and dynamic networks.Critical drivers,indispensable and neutral nodes were enriched with TFs,while the enrichment of essential genes in different categories of nodes were varied in different types of networks.The above results have certain robustness against network structure perturbations,which will provide some insights on real-world control of biological networks and have potential application in network medicine.