Research On Modeling And Immune Algorithm Optimization Of Immunotherapy Bayesian Gene Regulatory Network

With the continuous development and research of biological science and technology,immunotherapy has been used in the treatment of complex diseases such as tumors and autoimmune diseases.The occurrence of disease is the result of the complex interaction between multiple genes.It often cannot accurately reflect the inherent laws of disease occurrence and metastasis that studying a single gene and its expression in isolation.Therefore,in the post-genomic era,it is one of the research hotspots in disease immunotherapy that studying the function and dynamic changes of the entire genome.And through gene regulation network modeling,it is helpful to understand the generation process and regulatory relationship of disease-related genes and their products in the immune system,and the principles and processes of disease occurrence,so as to realize the overall understanding of disease.Gene regulation network modeling can contribute to the discovery,diagnosis,control and treatment of disease treatment,and the screening and development of drugs.What's more,it has an important impact on disease prevention and human health.Therefore,it has become the basis and hotspot in the research of disease immunotherapy that gene regulatory networks were constructed from gene expression data to discover regulatory relationships between genes and find targeted genes.However,due to the many types of gene regulatory network construction models and complicated calculations,more improvements and optimization studies are needed.After comparing the existing network models,the Bayesian network model is selected as the basic model for gene regulation network modeling in this paper.And through the research of immune algorithm,it was found that immune algorithm can better solve the NP-hard problem.Inspired by this,the immune algorithm is improved,and a genetic regulation network construction algorithm for learning the Bayesian network structure with the improved immune algorithm is proposed,which improves the accuracy of constructing the network structure.(1)An improved immune algorithm for immunotherapy gene regulation network is proposed.Compared with the general immune algorithm,the improved algorithm improves the processing method of memory cells,and proposes strategies for dynamic extraction and vaccination,which improves the stability and convergence speed of the algorithm and enables the algorithm to better converge to the global optimal solution.On the basis of solving the classic NP problem,by comparing and analyzing the results of the general immune algorithm and the immune algorithm based on the shortest path vaccine,the effectiveness of the improved immune algorithm is verified.(2)This paper proposes a genetic control network structure learning algorithm based on the improved immune algorithm,which reduces the amount of calculation and improves the accuracyof the network structure.On the framework of learning Bayesian network structure based on scoring search method,the improved immune algorithm is used as the network structure search algorithm,and the BIC scoring function is used as the fitness function in the algorithm to judge the quality of the network structure.By comparing the running results of the classic K2 algorithm,the hill climbing algorithm and network structure learning algorithm based on general immune algorithm to build a standard network with a known structure,it proves that the algorithm has higher accuracy.(3)The proposed network structure learning algorithm is used to construct a differentially expressed gene regulatory network for cancer.The paper selects breast cancer patient sample data from the GEO database.After screening the differentially expressed genes,the proposed algorithm was used to construct a differentially expressed gene regulatory network for breast cancer.Through comparative analysis of related research results,the node genes at important positions in the network are analyzed to find breast cancer-related genes,which proves that the gene regulatory network constructed by the algorithm has certain application value for the research of disease immunotherapy.