Multi-Omics and In Silico Drug Discovery Approaches to Identify Novel Biomarkers of Early Metastasis in Melanoma

Abstract: 

Melanoma is one of the most aggressive skin cancers, with early metastasis being the primary cause of patient mortality. Despite advances in targeted and immunotherapies, reliable biomarkers that predict early metastatic potential remain unclear. This study sought to identify novel genes associated with melanoma metastasis through an integrative bioinformatics pipeline combining transcriptomic, network, pathway, and drug discovery analyses. Differential expression analysis of the GSE7553 dataset identified 25 significant DEGs (p < 0.05, |log2FC| > 2), among which keratin and small proline-rich protein (SPRR) family members (KRT6A, KRT16, KRT17, SPRR1A, and SPRR3) were consistently upregulated in metastatic samples. Protein-protein interaction (PPI) networks and pathway enrichment analyses highlighted their central roles in keratinisation, cornified envelope formation, and developmental biology pathways, suggesting dysregulation of epidermal differentiation as a driver of tumour progression. Cross-validation with independent datasets and OncoPrint analysis confirmed their clinical relevance, while Kaplan-Meier survival curves linked high KRT17 and SPRR3 expression with poor prognosis. Finally, all five candidate genes were evaluated using Drugnome AI to determine their raw scores and percentile rankings for druggability as small-molecule targets in an oncogenic setting. Based on this analysis, the two most promising druggable candidates were identified and prioritised for subsequent wet-lab validation.