Integrative genomic and sequence-to-function analyses prioritize TNRC6B- linked regulatory variants in erysipelas susceptibility

Abstract Erysipelas is an acute skin and soft-tissue infection predominantly caused by streptococci, yet the host genetic architecture and non-coding regulatory mechanisms underlying susceptibility remain poorly characterized. We integrated FinnGen R11 erysipelas GWAS data (24,825 cases and 399,149 controls) with GTEx v8 and multiple functional genomic resources. Cross-tissue and tissue-specific TWAS, MAGMA, and Polygenic Priority Score (PoPS) analyses were combined with regional association testing, functional annotation, eQTL/sQTL and colocalization analyses, ENCODE chromatin annotations, rE2G/ABC enhancer-to-gene mapping, and AlphaGenome, Borzoi, and Enformer sequence-to-function models. TNRC6B was the only gene remaining significant after FDR correction in UTMOST (P = 1.21 × 10⁻⁵, FDR = 0.045) and was further supported by FUSION, MAGMA (Z = 3.4697, P = 2.61 × 10⁻⁴), and PoPS (score = 0.7712; genome-wide rank 324/18,383; local rank 2/31). No regional variant reached genome-wide significance. Integrated regulatory analyses prioritized rs4821940 and rs6001827 as Tier A candidates, whereas the 3′UTR variant rs139915 was classified as Tier B. Sequence-to-function models provided context-dependent support, although effect directions varied across models and tissues. Evidence-module integration, leave-one-module-out analyses, and 5,000 random weight perturbations consistently ranked rs4821940, rs6001827, and rs139915 among the top three candidates. These findings prioritize TNRC6B and linked regulatory variants for further investigation but do not establish causal effects.