rs118013985 (EIF3E): Male Pattern Baldness Locus

Key takeaways

• Over 250 independent genetic loci linked to male pattern baldness were identified in a genome-wide study of 52,874 men, the largest such study at the time of publication.
• Common genetic variants together explain roughly 50% of the variation in baldness susceptibility among men.
• A polygenic score built from these variants distinguishes no-hair-loss from severe hair loss with an AUC of 0.78.
• Men in the top 10% of polygenic scores had a 58% rate of moderate-to-severe hair loss, compared to 14% among those below the median score.

What the research says A genome-wide association study (GWAS) of 52,874 male UK Biobank participants (ages 40-69) identified over 250 independent genetic loci associated with severe male pattern baldness at genome-wide significance (P<5x10-8). The single-nucleotide polymorphism (SNP)-based heritability of baldness from common variants is estimated at around 50%, consistent with twin-study estimates of approximately 80% total heritability. A polygenic prediction algorithm built from common variants achieved an area under the curve (AUC) of 0.78 (sensitivity 0.74, specificity 0.69, positive predictive value 59%, negative predictive value 82%) in distinguishing individuals with no hair loss from those with severe hair loss.

Reported associations

• Male pattern baldness (severe hair loss): Among more than 250 independent genetic loci identified at genome-wide significance (P<5x10-8) in a GWAS of 52,874 UK Biobank men; this is the largest published GWAS of this trait at the time of publication, and the polygenic score derived from these loci explained substantially more variance than prior smaller predictors.

Evidence quality This study (n=52,874) is over four times the size of the previously largest meta-analytic GWAS of male pattern baldness (n=12,806, which identified 8 independent loci; the strongest individual signal at that time was an AR-region SNP with an odds ratio of 2.2). Loci were identified at genome-wide significance (P<5x10-8) and validated using a two-sample design: 40,000 participants contributed to the discovery GWAS and a separate 12,000 served as the validation target. Prior multi-SNP predictors achieved AUC values of 0.69-0.74 even when age was included as a covariate; the current multi-locus approach achieves AUC 0.78 without age, suggesting genuine improvement. The sample is limited to UK Biobank participants aged 40-69 years, which may reduce generalizability to younger men or populations of non-European ancestry. No conflicting findings are reported for the loci identified in this study.

Lifestyle considerations No lifestyle considerations on file for this variant.