Fig 2: EIF3B is a poor prognostic marker in melanoma. (A) Venn diagram shows the intersection of Cluster3 specific upregulated genes and the genes which are essential to melanoma human cell line growth using the DepMap database. (B) Kaplan–Meier analysis showing the association between the expression of EIF3B and SKCM patient overall survival (OS) in TCGA cohort (B) and GEO cohorts (C), the Log-rank test was used to access the statistical differences. The differences of EIF3B expression between different subtypes in TCGA dataset (D) and GEO datasets (E), the Anova test was used to calculate the statistical differences. (F) RT-qPCR and Western blot were used to verify the efficiency of EIF3B knockdown in SK-MEL-28 and A375 cell line. ‘*’ indicates p-value ≤ 0.05.

Fig 3: High expression of EIF3B correlates with low immune cells infiltration and predict the clinical benefit of ICB. (A) Top twenty enriched pathways in high and low EIF3B expression group were identified by GSVA analysis. (B) The graph showing the correlation of EIF3B expression and immune cell signature based on ssGSEA analysis output. The Spearman rank correlation test was used to calculate the correlation coefficients. (C) Correlations among EIF3B, LAG3, PD-1, CD8A, GZMB CTLA4, BTLA and IFNG levels in TCGA SKCM dataset. The Pearson correlation test was used to calculate the correlation coefficients. (D) The submap analysis shows that the lower EIF3B group in the GEO cohort is more sensitive to anti-pd-1 treatment. P-values were obtained after being adjusted by the Boferroni method.