The role of rare variants in male-pattern hair loss: Analysis of whole-exome sequencing data in the UK Biobank

Male-pattern hair loss (MPHL) is a highly heritable and prevalent form of hair loss. While genome-wide association studies (GWAS) have provided insights into the contribution of common genetic variants to MPHL etiology, the relevance of rare variants is unclear. In this study, we tested for a contribution of rare coding variants to MPHL etiology based on exome data of 72,469 men from the UK Biobank. We performed SKAT-O and GenRisk gene-based association analyses as well as single-variant tests, followed by interpretation, risk prediction modelling and enrichment testing. Men were classified as cases or controls based on self-reported hair loss pattern (1-unaffected, 2-frontotemporal balding, 3-frontotemporal and vertex balding, 4-baldness of the top of the scalp). We used three classification schemes: (i) controls (pattern 1) were compared to cases (patterns 2-4), (ii) supercontrols (pattern 1, age ≥ 60) were compared to severe cases (pattern 4, age < 60) and (iii) controls (patterns 1-2) were compared to cases (patterns 3-4), addressing a plausible self-report misclassification between patterns 1 and 2. The overall contribution of rare variants to MPHL based on risk prediction was minimal. Nevertheless, our analyses identified significant associations of MPHL and rare variants in 125 genes and 2 single variants. Associated genes include previously implicated candidate genes (e.g. WNT10A) and novel candidate genes at and beyond GWAS risk loci (e.g. CDH1, SPINK5). We further observe an enrichment of genes causative for monogenic trichoses. In summary, while our data point to a low contribution of rare coding variants to MPHL, they provide evidence for candidate genes and variants at and beyond known risk loci and suggest an association between monogenic trichoses and the common MPHL phenotype.