SUMMARY
We developed four polygenic risk scores (PGS) for primary open-angle glaucoma (POAG), which is the leading cause of irreversible blindness worldwide and remains undiagnosed in over half of patients. We constructed two genome-wide PGS using genome wide association study from African ancestry subjects: 1) the Primary Open-Angle African Ancestry Glaucoma Genetics (POAAGG) study (N = 7,031; POAAGG PGS) and an African ancestry GWAS (N = 11,275; MEGA PGS). We also derived two selected loci PGS from six multi-ancestry glaucoma GWAS and weighted these scores using African ancestry effect sizes (PGS616 and PGS526). In an independent training cohort (N = 271), the curated loci-based score PGS526 demonstrated the strongest standalone performance (mean AUC = 0.668), outperforming the genome-wide PGS constructed using PRS-CS. Integration with baseline demographic features (age and gender) further improved prediction, with the base + PGS616 model achieving a peak AUC of 0.806 with support vector machine model. Clinical enrichment in an independent suspect cohort (N = 1,013) showed that higher predicted genetic risk was significantly associated with elevated intraocular pressure, larger cup-to-disc ratio, and thinner retinal nerve fiber layer, which are all POAG diagnostic features. Leveraging inter-eye asymmetry, PGS further enhanced early disease discrimination, improving AUC from 0.823 to 0.862 for ΔIOP, from 0.769 to 0.817 for ΔCDR, and from 0.790 to 0.831 for ΔRNFL. These results demonstrate that PGS enhances prediction in phenotype-rich settings and enables accurate risk stratification in deep phenotype-limited cohorts, supporting earlier glaucoma detection.
HIGHLIGHTS
Ancestry-matched polygenic scores improve POAG risk prediction in African ancestry populations
Curated loci-based PGS outperform genome-wide scores in machine learning models
Genetic risk integrates with minimal demographics to achieve AUC up to 0.806
Predicted risk aligns with optic nerve damage and inter-eye asymmetry in suspects
Competing Interest Statement
The authors have declared no competing interest.
Funding Statement
This study was supported by the National Eye Institute, National Institutes of Health, through a Research Project Grant (R01EY023557) and a Vision Research Core Grant (P30EY001583). Additional funding was provided by the F.M. Kirby Foundation, Research to Prevent Blindness, the University of Pennsylvania Hospital Board of Women Visitors, and the Paul and Evanina Bell Mackall Foundation Trust. Institutional support was received from the Department of Ophthalmology at the Perelman School of Medicine and the VA Hospital in Philadelphia, Pennsylvania.
Author Declarations
I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.
Yes
The details of the IRB/oversight body that provided approval or exemption for the research described are given below:
The Institutional Review Board of the University of Pennsylvania gave ethical approval for this work (protocol #812036). All participants provided written informed consent.
I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.
Yes
I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).
Yes
I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.
Yes
Footnotes
↵# Lead contact
This revised version of the manuscript includes substantial methodological, structural, and analytical updates to improve clarity, rigor, and biological validation. First, the genetic modeling framework has been refined. The prior version emphasized curated Genetic Risk Scores GRS 1 to 4 derived from selected African ancestry loci. The revised version introduces four clearly defined Polygenic Risk Scores constructed using two complementary strategies: genome wide PRS using PRS CS from African ancestry GWAS datasets, and loci based PGS derived from multi ancestry glaucoma GWAS and re weighted using African ancestry effect sizes. The manuscript now provides expanded methodological detail regarding SNP selection, linkage disequilibrium handling, normalization, and weighting procedures. Second, cohort structure and sample definitions were clarified and strengthened. The training cohort has been more explicitly separated from the PGS derivation datasets to avoid overlap. The revised version clearly distinguishes the PGS generation cohort, the clean training cohort, and the independent suspect validation cohort. Sample sizes and phenotype thresholds were standardized and described more transparently. Third, model evaluation has been expanded. The revised manuscript systematically compares Random Forest, Support Vector Machine, and Multilayer Perceptron classifiers across consistent feature sets and reports comprehensive performance metrics. Cross model averaging and structured benchmarking were added to improve interpretability. Fourth, biological validation analyses were substantially strengthened. The updated version includes enrichment testing in an independent suspect cohort using continuous predicted risk scores, correlation analyses with intraocular pressure, cup to disc ratio, and retinal nerve fiber layer thickness, and formal statistical testing. In addition, inter eye asymmetry analysis was introduced and evaluated as an independent validation framework, demonstrating improved early disease discrimination when genetic risk is integrated. Finally, the Discussion and Limitations sections were revised to better emphasize ancestry matched genetic modeling, generalizability considerations, interpretability trade offs, and translational implications for equitable glaucoma screening. These revisions substantially enhance methodological transparency, biological grounding, and clinical relevance.
Data Availability
The data underlying this study are not publicly available due to participant privacy and IRB restrictions. De-identified data may be made available from the corresponding author upon reasonable request and completion of a data use agreement. Code is available from the corresponding author upon request.
