Integration of scRNA-seq and Bulk-seq to Analyse the Infiltration of Monocytes in Pancreatic Cancer and Establish a Molecular Risk Model

Abstract Background Many researches have confirmed that immunotherapy of tumor immune microenvironment is necessary. In pancreatic cancer, monocytes play an important role in poor prognosis, but the mechanism and prognosis prediction methods are unclear. Methods CIBERSORT was used to identify cellular immune score and evaluate the effect of each immune cell on prognosis. The gene modules related to monocytes were obtained by weighted correlation network analysis through WGCNA package. Consensus clustering was used to sort prognostic genes. The regression signature was generated by LASSO Cox analysis and verified by Cox analysis. The ssGSEA and TIDE algorithms were used to predict immune status and sensitivity to ICB. Finally, the expression levels of each gene were verified at tissue level and single-cell level. Results High infiltration of monocytes suggests poor prognosis of pancreatic cancer. 262 genes were significantly differentially expressed and prognostic after WGCNA analysis and cluster typing. The related 6 genes prognostic signature established by LASSO Cox analysis was verified to be an independent prognostic factor. The high-risk group had high infiltration of monocytes in the immune microenvironment and was more sensitive to ICBs. At the tissue level, all genes were highly expressed in cancer tissues. At the single-cell level, MET and MYEOV were significantly higher in malignant cells and lower in monocytes. Conclusions High infiltration of monocytes affects poor prognosis of pancreatic cancer, suggesting that the immune microenvironment has a certain research prospect for treatment of pancreatic cancer. The monocyte-related genes signature can accurately assess the prognostic risk of pancreatic cancer.