Construction of a prognostic model for ovarian cancer based on a comprehensive bioinformatics analysis of cuproptosis-associated long non-coding RNA signatures

Abstract

Ovarian cancer (OCa) is a common malignancy in women, and the role of cuproptosis and its related genes in OCa is unclear. Using the GSE14407 dataset, we analyzed the expression and correlation of cuproptosis-related genes (CRGs) between tumor and normal groups. From the TCGA-OV dataset, we identified 20 cuproptosis-related long non-coding RNAs (CuLncs) associated with patient survival through univariate Cox analysis. OCa patients were divided into early-stage and late-stage groups to analyze CuLncs expression. Cluster analysis classified patients into two clusters, with Cluster1 having a poorer prognosis. Significant differences in “Lymphatic Invasion” and “Cancer status” were observed between clusters. Seven CRGs showed significant expression differences, validated using the human protein atlas (HPA) databases. Immune analysis revealed a higher ImmuneScore in Cluster1. GSEA identified associated signaling pathways. LASSO regression included 11 CuLncs to construct and validate a survival prediction model, classifying patients into high-risk and low-risk groups. Correlations between riskScore, Cluster phenotype, ImmuneScore, and immune cell infiltration were explored. Cell experiments showed that knocking down AC023644.1 decreases OCa cell viability. In conclusion, we constructed an accurate prognostic model for OCa based on 11 CuLncs, providing a basis for prognosis assessment and potential immunotherapy targets.

Keywords

• Cuproptosis
• Prognostic model
• Long non-coding RNAs signature
• Patient survival
• Immune infiltration