Research On Gene Regulatory Network Reconstruction And Identification Of Essential Genes And Their Application In Leukemia

Leukemia is a common cancer associated with hematopoietic stem cells,it has high recurrence rate and complicated mechanism,and this disease can cause a variety of complications.Targeted therapy with drugs can control the development of the disease,and it has little effect on normal cells,but this measure can only treat the cancer sites which have been identified.Since the pathogenesis of leukemia is not yet known,many potential disease-related genes need to be explored.The method of traditional biological experiments to identify essential genes is inefficient,however the use of bioinformatics data,combined with computer science,complex network science,and data mining techniques,can avoid this shortcoming.In order to find potential therapeutic targets of leukemia,we mainly analyze leukemia from gene regulatory network,gene co-expression network and network controllability,and then find out key genes related with leukemia.The specific work is summarized as follows:1)Firstly,in order to construct a more precise gene regulation network,we propose a new gene network reconstruction algorithm based on ‘deepwalk',we can use new and effective biological information data set,in addition we can avoid the use of high-dimensional sparse features.Compared with the current model,we verify the feasibility and effectiveness of our method.In this paper,the leukemia gene regulatory network was constructed based on the above algorithm,and 10 key genes were found based on the topology of the network.2)Weighted Gene Co-Expression Network Analysis(WGCNA)is another important method to find essential genes which are widely used in the association analysis of disease and trait or gene.It assumes that the gene network obey a scale-free distribution,so that the network constructed is more biologically significant.The experimental results show that the WGCNA method can find biologically meaningful modules.Finally,we find 223 high-connectivity genes in three modules,and some of the genes were consistent with the literature reports,indicating that these genes are important in the pathogenesis of leukemia.3)The structural controllability analysis of the network is a dynamic analysis method.The maximum flow matching algorithm based on the graph is used to find the minimum driving node set of the leukemia gene network,and then we divide the node into two kinds based on the change of the size of minimum driving node set after deleting the node,we named it as ‘neutral' node and ‘dispensable' node.After the enrichment analysis of the two types of nodes on important biological data sets and the robustness analysis of the two types of nodes,we find that the ‘neutral' nodes have a greater impact on the network,so the neutral nodes have more potential research value.