Leveraging expression quantitative trait loci information in single-cell resolution to identify cell-specific genes for Basal cell carcinoma

Abstract

Background Basal cell carcinoma (BCC), the most common skin cancer, is driven by UV-induced DNA damage and shaped by immune surveillance. Although GWAS has identified over 140 risk loci, their cell-type-specific effects remain obscured by tissue-level averaging. Methods We integrated BCC GWAS summary statistics from a UK-based cohort (17,416 cases, 375,455 controls) with single-cell expression quantitative trait locus (sc-eQTL) data from 12 immune cell types in the OneK1K resource. Using the OTTERS framework combined with ACAT-O, we performed single-cell transcriptome-wide association analysis (scTWAS); bulk TWAS using GTEx whole blood served as a conventional tissue-averaged benchmark for comparison, rather than a definitive gold standard. Functional enrichment was conducted via Gene Ontology (GO). Results Bulk TWAS using GTEx whole blood identified 35 BCC-associated genes (FDR < 0.05), including MC1R and CASP8 . In contrast, single-cell transcriptome-wide association study (scTWAS) across 12 OneK1K cell types revealed 207 non-redundant susceptibility genes, predominantly in CD4 ET , MONO C , and B IN . Functional enrichment uncovered cell-type-specific programs: MHC class II antigen presentation (CD4 ET ), PRR-mediated innate immunity (MONO C ), and pro-inflammatory secretion (B IN )—all absent in bulk results. Conclusion Our exploratory scTWAS using healthy donor PBMC-derived eQTLs uncovered cell-type-specific BCC associations missed by bulk analyses, providing hypothesis-generating insights into immune-related genetic effects in BCC. This underscores how single-cell resolution can overcome signal dilution from cellular heterogeneity, though validation in tumor-derived immune populations is warranted.