Application Of Network Clustering Algorithm Based On Exponential Family Graph Model In Single-cell Sequencing Data

In recent years,with the in-depth study of tumor diseases,many new cancer subtypes have been found,and the classification of local tumor types according to traditional methods cannot meet the current clinical needs.People gradually realize the importance of studying tumor heterogeneity at the single-cell level.The emergence of single-cell RNA-Seq technology makes the study of heterogeneity possible at the single-cell level.It can provide the expression information of nearly ten thousand genes in a single cell,and provide a powerful tool for understanding the relationship between genotype and phenotype,and comprehensively reveal the heterogeneity of gene expression between cells.However,the genetic network is always heterogeneous,and the traditional statistical network model is not enough to capture this characteristic,which results in the network information in the single-cell sequencing data cannot be fully utilized,resulting in the waste of genetic information.In this article,we propose a hybrid network clustering algorithm based on exponential family figure(Mixture ERGM),and the model algorithm is applied to single cell sequencing data,considering because any clustering algorithm can get the corresponding cell clusters,so the results of clustering analysis of WGCNA enrichment analysis and inspection,at the same time to the RP analysis of clustering results,point out the differentially expressed genes in gene expression data,using internal high expressed genes of different cell types in the function of the module to participate in the close relationship between cell types,through enrichment analysis of differentially expressed genes and their related proteins and biological pathway research,the objective of revealing the heterogeneity of cells was achieved.