Identification of macrophage-related genes in bladder cancer patients using single-cell sequencing and construction of a prognostic model.

Abstract

Single-cell sequencing is an emerging technology that can effectively identify cell types in tumors. In the tumor microenvironment of bladder cancer, macrophages play a crucial role in invasion and immune escape. This study aimed to assess the expression of macrophage-related genes (MRGs) in the tumor microenvironment of bladder cancer patients and construct a prognostic model based on MRGs using bioinformatics methods.

Methods: Single-cell sequencing data from bladder cancer patients was downloaded from the GEO. After quality control and cell type identification, macrophages in the samples were extracted for re-clustering. Feature genes were then identified, and MRGs were assessed. Genetic data from TCGA database bladder cancer patients was also downloaded and organized. The intersection of MRGs and the TCGA gene set was determined. Clinical information was connected with this intersection, and the data was divided into training and validation sets. The training set was used for model construction and the validation set for model verification. A prognostic model based on MRGs was built using the LASSO algorithm and Cox regression. Patients were divided into high-risk and low-risk groups based on their prognostic features, and survival information in the training and validation sets was observed. The predictive ability of the model was assessed using a ROC curve, followed by a calibration plot to predict 1-, 3-, and 5-year survival rates.

Results: Four cell types were identified, and after extracting macrophages, three cell subgroups were clustered, resulting in 1,078 feature genes. The top 100 feature genes from each macrophage subgroup were extracted and intersected with TCGA expressed genes to construct the model. A risk prediction model composed of CD74, METRN, PTPRR, and CDC42EP5 was obtained. The survival and ROC curves showed that this model had good predictive ability. A calibration curve also demonstrated good prognostic ability for patients.

Conclusion: This study, based on single-cell data, TCGA data, and clinical information, constructed an MRG-based prognostic model for bladder cancer using multi-omics methods. This model has good accuracy and reliability in predicting the survival and prognosis of patients with bladder cancer, providing a reference for understanding the interaction between MRGs and bladder cancer.