Identification of biomarkers associated with melanoma based on bioinformatics analysis

Abstract

Background: Melanoma is an aggressive tumor caused by melanocytes characterized by a poor prognosis at the stage of metastasis. Therefore, it is crucial to identify biomarkers for progression and prognosis for the treatment of melanoma. Objectives: The study aimed to identify the specific genes related to the prognosis of melanoma utilizing bioinformatics analyses. Methods: Differentially expressed genes (DEGs) between melanoma tissues and normal tissues were screened from the GSE 3,189 gene expression profile of Gene Expression Omnibus (GEO). A co-expression network was constructed by weighted correlation network analysis (WGCNA). Functional enrichment analysis for DEGs was performed. The risk prognostic model and nomogram predictive model were established utilizing the least absolute shrinkage and selection operator (LASSO) and Cox regression analysis. Results: Using the differential expression analysis and WGCNA, 1,408 DEGs were screened between melanoma tissues and normal tissues. Functional enrichment analysis proved that these genes primarily participated in the cell cycle and mitotic phase regulation in cancer. In addition, 17 optimal DEGs were identified for constructing the risk score prognostic model. Cox regression analysis further revealed that ectonucleotide pyrophosphatase/phosphodiesterase 4 (ENPP4) and FGR proto-oncogene, Src family tyrosine kinase (FGR) were the key genes significantly associated with survival. A nomogram prediction model was established for individual survival probability by integrating pathological T/N/M stage, age, ENPP4, and FGR. High FGR or ENPP4 expression indicated a better prognosis in melanoma patients. Conclusion: This study identified FGR and ENPP4 as potentially useful prognostic biomarkers for melanoma. The corresponding risk score prognostic model and nomogram may be a reliable tool for predicting the prognosis of melanoma patients.