Construction Of Breast Cancer Typing And Prognosis Model Based On Single-cell Sequencing

Breast cancer is the most common malignant tumor in women and is a complex disease with high heterogeneity.Current clinical treatment,mostly using surgery and chemotherapy,is highly effective in early-stage patients,but less effective in some intermediate and late-stage patients.Single-cell sequencing,which provides gene expression profiles of individual cells and helps to reveal intercellular heterogeneity,has become an important tool to explore cellular heterogeneity and complex biological processes.To further elucidate the heterogeneity and complexity of breast cancer,this study integrated single-cell sequencing data and transcriptomic data to investigate the important role of cell-cell interactions and related cytokines in the development of breast cancer at the cellular and genetic levels,with the aim of providing data for personalized treatment of breast cancer.Multiple breast cancer single cell datasets were collected from the public database of GEO(Gene Expression Omnibus),and finally 100,064 cellular data were screened.First,the cells were clustered using PCA method to obtain 18 clusters of cells,and the clustered cells were annotated into 9 cell types,including epithelial cells,endothelial cells,tissue stem cells,fibroblasts,monocytes,macrophages,NK cells,B cells and T cells.Further analysis of the intercellular communication network by cell communication identified 68 significant ligand-receptor pairs in the nine cell types.Functional enrichment analysis showed that these ligand-receptor pairs were mainly distributed in 24 cellular signaling pathways.Interestingly,we found that macrophages can communicate with endothelial and epithelial cells in VEGF,MIF,SEMA3 and MK signaling pathways,which may be relevant to the induction of tumor cell survival and migration processes.Second,we screened 21 candidate genes highly associated with macrophages through a weighted gene co-expression network,which were significantly enriched in MHC antigen processing and exogenous peptide antigen presentation pathways and may be involved in immune response processes in tumors.To understand the potential roles of these candidate genes in breast cancer,nine genes that were significantly and positively associated with the M1 phenotype of macrophages were screened in combination with TCGA(The Cancer Genome Atlas)transcriptome data and their roles were explored in terms of both prognosis and staging.Based on a combination of various machine learning techniques,we identified an optimal prognostic evaluation model using genes related to macrophages(CD74,SAMHD1,HLA.DPB1,and HLA.DMB).Two molecular subtypes of breast cancer,Cluster1 and Cluster2,were identified using a principal component analysis algorithm,which in turn compared the tumor microenvironment and differences in treatment response between the two subtypes.The results indicated that Cluster1 showed a better prognosis,but Cluster2 had a higher level of immune infiltration and most of the immune checkpoint markers tended to be overexpressed in Cluster2 and more sensitive to some common chemotherapeutic agents,showing a good prospect for single and combination therapy.Finally,based on the above findings,BRCAGAP(BRCA Gene Analysis Platform,http://www.tmliang.cn/brca/),a breast cancer-related database,was developed.BRCAGAP integrated breast cancer-related multi-omics data and provided an integrated analysis platform.BRCAGAP allowed users to perform a variety of data analysis such as expression analysis,immune signature mining and drug prediction,which can provide a powerful data analysis platform for researchers.In summary,this study mapped breast cancer single cell profiles based on single cell sequencing data,explored intercellular communication patterns during breast cancer development,and identified 21 macrophage-specific genes with predictive value and 9 genes with significant positive correlation to the M1 phenotype of macrophages in combination with transcriptome sequencing data.Then a prognostic model was constructed,two cancer subtypes with immune microenvironmental heterogeneity were identified,and a breast cancer-related database,BRCAGAP,was developed.These results provide a basis for further exploration of the cellular mechanisms involved in the development of breast cancer and new ideas for the precise diagnosis and treatment of breast cancer,which can provide data reference and theoretical support for the development of new personalized immunotherapy targets and anticancer drugs.