Abstract
Ferroptosis plays a critical role in different types of cancers, but the prognostic impact of ferroptosis in cutaneous melanoma remains lacking. Therefore, ferroptosis-related genes (FRGs) were firstly obtained from the FerrDb database and the differentially expressed FRGs were identified by the "limma" algorithm. Next, the prognostic differentially expressed FRGs were screened out by univariate Cox regression, which were subsequently used to cluster melanomas into two subtypes (clusters A and B). Besides, the Boruta algorithm and principal component analysis (PCA) were performed to build a 15-FRGs indicator, which can robustly predict patients' overall survival (OS) and be considered as an independent prognostic factor in melanoma. The melanoma patients were further divided into high- and low-FRGs score groups. The high score group have a good prognosis, with higher T cell immune infiltrating and lower mutation frequencies in NRAS, KRAS, and NF1. Finally, we discovered that many immune processes and several chemotherapy drugs were closely associated with FRGs score. Thus, our study provides a novel ferroptosis-associated classifier and indicator to predict the prognosis of melanoma. Besides, we identified several potential chemotherapy drugs to induce ferroptosis and could supply additional effective treatments.