458 A focused analysis of genome-wide associations of dermatologic conditions in the UK biobank

Genome-wide association studies (GWAS) link common variants to a trait of interest. Decreasing sequencing costs enable large-scale efforts such as the UK Biobank genotype and phenotype ∼500,000 individuals for over 7,000 traits. However, due multiple testing corrections approximately 200 billion tests, many clinically statistically significant associations are unappreciated. By re-analyzing data 13 dermatologic conditions, we hypothesized would identify well-established new genetic that can shed insight into disease prediction novel therapeutic targets. After filtering phenotypes conditions with at least 500 cases, identified traits this analysis: malignant melanoma, melanoma in situ, squamous cell carcinoma, basal actinic keratosis, seborrheic psoriasis, lichen planus, systemic lupus erythematosus, hyperhidrosis, pilonidal cyst, sebaceous lipoma. We filtered 166,373,136 Benjamini-Hochberg false discovery rate 5% yield 447 sentinel (the most variant within haplotype block (250Kb R2>0.1)). These were enriched protein-coding (p<1e-5). The validated previously including genes affecting skin pigmentation—Tyrosinase (TYR), Melanoma Antigen AIM1 (SLC45A2/AIM1), Melanocortin 1 Receptor (MC1R)—with melanoma. Novel LDL Related 3 (LRP3) lipomas Phospholipase C Delta (PLCD1) cysts also identified. When analyzing all variants, noted higher CADD score (predictor deleteriousness) compared random (p=0.0037). availability databases redefining modern GWAS. Our focused analysis has numerous hundreds variety conditions.