Conduction and validation of a novel prognostic signature in cervical cancer based on the necroptosis characteristic genes via integrating of multiomics data

Abstract

The significance of necroptosis in recurrent or metastatic cervical cancer remains unclear. In this study, we utilized various bioinformatics methods to analyze the cancer genome atlas (TCGA) data, gene chip and the single-cell RNA-sequencing (scRNA seq) data. And a necroptosis-related genes signature for prognostic assessment of patients with cervical cancer was constructed successfully. Survival analysis, receiver operating characteristic (ROC) curve, the support vector machine recursive feature elimination (SVM-RFE) algorithm and random forest analysis were performed to validate this signature. Patients in TCGA-CESC cohort were grouped into "high-necroptosis score (H-NCPS)" vs "low-necroptosis score (L-NCPS)" subgroups based on the median of necroptosis score of each patient. Analyses of the tumor microenvironment manifested "H-NCPS" patients associated with lower degree of immune infiltration. Through the utilization of survival analysis, cell communication, and Gene Set Enrichment Analysis (GSEA), PGK1 was determined to be the pivotal gene within the 9-gene signature associated with necroptosis. The high expression of PGK1 in cervical cancer cells was confirmed through the utilization of quantitative real-time polymerase chain reaction (RT-qPCR) and the human protein atlas (HPA). In the interim, PGK1 prompted the transition of M1 macrophages to M2 macrophages and influenced the occurrence and development of necroptosis. In conclusion, the 9-gene signature developed from necroptosis-related genes has shown significant predictive capabilities for the prognosis of cervical cancer, offered valuable guidance for individualized and targeted treatment approaches for patients.

Keywords: Cell communication; Cervical cancer; Machine learning; Necroptosis; PGK1; Prognostic signature; Pseudotime analysis; WGCNA.