Genome-Wide Association Study of Corneal Dystrophy Uncovers Novel Risk Loci and Enables Improved Polygenic Prediction...

To identify risk loci for hereditary corneal dystrophy and improve a polygenic prediction model for Fuchs endothelial corneal dystrophy (FECD).

We conducted a meta-analysis of 7316 Europeans with hereditary corneal dystrophy (including FECD and related conditions) and 1,588,467 controls from the UK Biobank, All of Us, FinnGen, and Million Veteran Program (MVP), assessing phenotype validity via genetic correlation. Polygenic risk scores (PRS) were derived using SBayesRC and validated in two independent European FECD cohorts. A genome-wide association study of African ancestry from MVP (455 cases and 121,154 controls) was used to develop African PRS. TCF4 CTG18.1 short tandem repeat (STR) lengths were inferred in All of Us to model STR expansion and PRS.

We identified 24 risk loci associated with corneal dystrophy, including 12 loci not previously reported in FECD studies. The optimized PRS achieved strong performance in two European FECD validation cohorts (area under the curve [AUC] = 0.83; 95% confidence interval, 0.82-0.84; DeLong's P = 7.04  ×  10-19), with individuals in the top decile showing 14‑fold and 19‑fold elevated risk. In All of Us, STR expansion (>40 repeats) was the key predictor of FECD risk (AUC = 0.89; odds ratio [OR] = 54), with minimal incremental value from PRS. STR expansion remained the primary risk driver across ancestries, whereas PRS provided modest independent predictive value for broader corneal dystrophy phenotypes in European and admixed American groups. Among individuals without large STR expansions, PRS was the only significant contributor in Europeans (OR = 1.37), whereas analyses of FECD non‑expansion carriers were underpowered.

These findings refine the genetic architecture of hereditary corneal dystrophy, improve risk prediction for FECD, and support a tiered strategy integrating STR expansion with PRS.