Uveal melanoma (UM) and cutaneous melanoma (CM) originate from the same melanocytic lineage, which motivates comparative genomic investigation to identify shared pathogenic mechanisms. The authors performed a database‑driven bioinformatic study to detect overlapping variant–disease associations (VDAs) between UM and CM and to explore potential molecular interactions that could explain shared or divergent clinical behavior.
The analysis used publicly accessible variant and disease association resources. VDAs were retrieved from the GWAS Catalog and DisGeNET, with variant validation performed using Ensembl. To evaluate potential molecular relationships among implicated gene products, the authors applied the STRING database to construct protein–protein interaction (PPI) networks. The study used only de‑identified, open data; no human subjects or new samples were involved.
From the initial queries, the authors identified 109 unique VDAs related to UM and 880 unique VDAs related to CM across the databases searched. These counts reflect the number of distinct variant–disease associations cataloged in the source databases for each melanoma subtype as retrieved by the study team.
Despite the larger number of VDAs tied to CM, only two VDAs were present in both UM and CM in the assembled dataset and across different ethnic groups. Those two shared single‑nucleotide variants are rs12203592 (mapped to the IRF4 gene) and rs12913832 (mapped to the HERC2 gene). The authors report these two variants as the only overlapping VDAs identified in their cross‑database comparison.
Using STRING, the authors evaluated interactions among proteins encoded by genes implicated in the VDA analysis. The PPI network highlighted that OCA2 and IRF4 directly interact with HERC2. These direct interactions suggest a mechanistic linkage between HERC2, pigmentation‑related proteins, and IRF4 that could be relevant to melanoma biology. The manuscript reports these interactions as part of the evidence connecting the shared variants to potential functional consequences.
The authors summarize differential prognostic associations for the two shared genes: HERC2 is reported as a poor prognostic factor in uveal melanoma, whereas IRF4 status is identified as a key prognostic indicator in both UM and CM. The paper emphasizes that characterization of IRF4 alleles could improve understanding of melanoma pathogenesis and aid development of disease‑specific approaches. The report does not present primary clinical outcome data; prognostic statements are drawn from the aggregated database associations and network context described in the manuscript.
This work is a bioinformatic review of existing public databases rather than a prospective clinical or experimental study. The manuscript is a preprint and has not been peer reviewed; the authors explicitly state that the findings should not be used to guide clinical practice. Because the analysis used publicly available, de‑identified data from GWAS Catalog, DisGeNET, and Ensembl, the authors report that formal ethics approval and informed consent were not required. Links to the three primary data sources are provided in the paper.
In this database‑driven analysis, two shared VDAs—rs12203592 (IRF4) and rs12913832 (HERC2)—were identified between uveal and cutaneous melanoma despite substantial divergence in the total number of VDAs cataloged for each tumor type. PPI analysis revealed direct interactions of OCA2 and IRF4 with HERC2, supporting a molecular connection among these proteins. The authors report that HERC2 associates with poorer prognosis in UM, while IRF4 emerges as an important prognostic marker across both UM and CM. The study frames these findings as a basis for further functional and clinical investigation, noting the preprint status and reliance on existing database annotations.
All facts and counts presented here are drawn from the cited preprint. Specific statistical metrics, cohort sizes, ethnicity‑level frequencies, or detailed outcome measures were not reported in the abstract and summary material provided in the source. Readers should consult the full preprint and the underlying databases for raw data and additional methodological detail.