Joint Analysis of the Interferon Gene Signatures and Cancer-Testis Gene Expression in Cutaneous Melanoma Patients

Introduction. Cutaneous melanoma currently lacks well-established molecular biomarkers for predicting immunotherapy response. Emerging candidates under investigation include interferon-stimulated gene (ISG) signatures and cancer-testis antigen (CTA) expression profiles. Aim. The study performed to characterize the interplay between ISG signatures and CTA expression patterns in cutaneous melanoma patients. Materials and Methods. The study utilized normalized whole-genome sequencing data comprising expression levels of 43,000 genes from 457 cutaneous melanoma patients, sourced from the publicly available University of California Santa Cruz (UCSC) dataset. Our analysis focused on rhabdoid-testicular CTA genes (n = 186) and interferon-dependent ISG genes (n = 66), the latter analyzed as both full and brief signatures. We performed agglomerative hierarchical clustering via Ward's method separately for ISG and CTA groups. Statistical evaluation of cluster interactions employed seven complementary measures: Pearson's chi-square test, lambda coefficient, contingency coefficient, phi coefficient, Goodman and Kruskal's tau, uncertainty coefficient, and column proportion analysis. Results. Analysis revealed four conserved ISG clusters across both datasets (two demonstrating high gene expression and two with low expression) showing strong correlation (λ = 0.666, p < 0.0001). For CTA genes, hierarchical clustering identified six primary clusters (two each of high, medium, and low expression genes) at the first level, which further differentiated into ten subclusters at the third clustering level. Initial comparison of first-level ISG and CTA clusters showed no significant association (p > 0.1). Evaluation of third-level CTA clusters against the brief ISG signature demonstrated a weak relationship (symmetric uncertainty coefficient = 0.031, p = 0.003). Only two third-level CTA clusters exhibited meaningful associations with ISG patterns: one characterized by minimal CTA expression coupled with high ISG activity, and another showing the inverse relationship of elevated CTA expression paired with low ISG signature. Conclusion. This study confirms our prior findings regarding the heterogeneous expression profile of CTA in cutaneous melanoma. The majority of CTA clusters demonstrated no significant association with ISG signatures. The identification of specific CTA-ISG expression patterns as predictive biomarkers for immunotherapy response in melanoma patients warrants deeper investigation.