Immune-related gene signature associates with immune landscape and predicts prognosis accurately in patients with skin cutaneous melanoma

Abstract

Skin cutaneous melanoma (SKCM) is the skin cancer that causes the highest number of deaths worldwide. There is growing evidence that the tumour immune microenvironment is associated with cancer prognosis, however, there is little research on the role of immune status in melanoma prognosis. In this study, data on patients with Skin cutaneous melanoma were downloaded from the GEO, TCGA, and GTEx databases. Genes associated with the immune pathway were screened from published papers and lncRNAs associated with them were identified. We performed immune microenvironment and functional enrichment analyses. The analysis was followed by applying univariate/multivariate Cox regression algorithms to finally identify three lncRNAs associated with the immune pathway for the construction of prognostic prediction models (CXCL10, RXRG, and SCG2). This stepwise downscaling method, which finally screens out prognostic factors and key genes and then uses them to build a risk model, has excellent predictive power. According to analyses of the model's reliability, it was able to differentiate the prognostic value and continued existence of Skin cutaneous melanoma patient populations more effectively. This study is an analysis of the immune pathway that leads lncRNAs in Skin cutaneous melanoma in an effort to open up new treatment avenues for Skin cutaneous melanoma.

Keywords: bioinformatics; immune pathway; prognosis; skin cutaneous melanoma; tumor environment.

FIGURE 1

DEG identification and analysis. (A)Volcano plot of DEGs in TCGA-SKCM, GSE15605, and GSE46517 datasets. (B) Venn diagram: Venn plots of DEGs in the TCGA-SKCM, GSE15605, and GSE46517 datasets, with the overlapped part representing the 16 common DEGs in the three data sets. (C) The 16 overlapping parts portray the DEGs that are shared by the three data sets.

FIGURE 2

Identification of immune-related SKCM subtypes in the TCGA cohort. (A) The heatmap depicts the consensus matrix at k = 2 in the TCGA group. (B)The cumulative distribution function (CDF) curves in consensus cluster analysis. Consensus scores for different subtype numbers (k = 2–9) are presented. (C)The stratification into three subtypes validated by t-SNE in TCGA cohorts. Each dot represents a single sample, and each color denotes a subtype. (D) Box figure:the 16 overlapped part represents the common DEGs in the two TCGA cohorts.(E) heatmap:the 16 overlapped part represents the common DEGs in the two TCGA cohorts.(F) Survival analysis of patients with the three diffuse glioma subtypes (IM-Hot and IM-Cool) in TCGA cohorts. The log-rank test was conducted to determine the significance of the differences.

FIGURE 3

Immune characteristics of the two subtypes in TCGA cohort. (A,B) The heatmap showing the abundance of immune-cell populations calculated by GSVA and Mcpcounter in the two subtypes. (C,D) Box plots show trends in the abundance of different immune cell infiltrations between the two subtypes in TCGA cohort.

FIGURE 4

GO and KEGG analysis results of genes. (A) The heatmap showing the DEGs in the two subtype TCGA cohorts. (B) Up-regulated genes in the IM-Hot group GO function enrich-ment results, and KEGG pathway enrichment results. (C) Up-regulated genes in the IM-Cool group GO function enrichment results, and KEGG pathway enrichment results.

FIGURE 5

Regression analysis. (A) Forest plots depicting the univariate Cox regression analysis. (B)multivariate Cox regression analysis.

FIGURE 6

Prognostic value of the proposed subtyping for SK CM. (A) Risk Score, and expression of 3-gene in TCGA training set. (B) KM survival curve distribution of 8-gene signature in training set. (C) ROC curve of 3-gene signature classification and AUC.