Research On Gene Regulatory Network In Complex Life Process

Since the completion of the Human Genome Project,modern life science research has came into a new age of systems biology.The birth of this new field represents that continuous exploration of new biological methods is about to begin.Scientists have gradually realized that the study of complex life processes is not limited to a single gene,but comprehensively and systematically to explore the expression regulation between genes,to reveal the operating mechanism of the entire life system,and finally cracks the secrets of life inheritance.Along with the rapid development of sequencing technology,the research results of various laboratories have produced a large amount of gene expression data.How to use the calculation method to mine biologically significant gene regulatory relationship and regulatory rules in these data has become one of the most challenging problems facing humanity in the post-genome era.Among the many models used to construct regulatory networks,the Bayesian Networks(BN)model has a solid theoretical foundation and flexible reasoning capabilities,and it is a powerful tool for building regulatory networks.With the development of sequencing technology,the advantages of the combination of Bayesian network model and new single-cell sequencing data will be more prominent.In this paper,taking the gene expression data for research object,based on the summary of the Bayesian method to construct the gene regulation network research,the existing scoring search algorithm has been improved,mainly completing two parts of the work:(1)A new method for screening key node genes in the network is proposed.Firstly,the co-expression network of all genes was constructed,and the importance of the genes themselves in the network was estimated by using Bayesian probability to obtain a quantitative table of gene importance.Secondly,the PageRank algorithm was used to estimate importance transfer caused by interaction between gene nodes in co-expression networks;finally,the key nodes are sorted by importance.The experimental results show that the proposed method has a significant improvement compared to existing key node screening algorithms.(2)Design a Bayesian network search algorithm based on information flow.In the process of constructing the regulatory network of key genes,first calculate the information flow between nodes,and construct the initial network,omitting the operation of reversing the edge in the search process to improve the search efficiency;then use the Tabu Search strategy to learn best network structure,setting contempt criteria that integrate prior knowledge to improve accuracy of the network.This paper implements these two algorithms and constructs a gene regulatory network on the single-cell transcriptome data of myocardial development.Finally,the results of this paper are compared with the results of other Bayesian network structure learning algorithms in terms of time cost and network accuracy,which proves the effectiveness of the algorithm designed in this paper.