Oral Presentation GENEMAPPERS 2026

Ensuring Glaucoma Polygenic Risk Scores Work for Everyone: Calibration Across Seven Underrepresented Populations (#6)

Ngoc-Quynh Le 1 2 , Guiyan Ni 3 , Regina Yu 1 , Owen M Siggs 4 5 , David A Mackey 6 , Jamie E Craig 7 , Alex W Hewitt 8 , Puya Gharahkhani 1 2 , Stuart MacGregor 1 2
  1. QIMR Berghofer Medical Research Institute, Herston, QLD, Australia
  2. Faculty of Medicine, University of Queensland, Brisbane, QLD, Australia
  3. Seonix Bio, Adelaide, South Australia, Australia
  4. Genomics and Inherited Disease Program, Garvan Institute of Medical Research, Sydney, New South Wales, Australia
  5. Faculty of Medicine and Health, University of New South Wales, Sydney, New South Wales, Australia
  6. Lions Eye Insitute, Perth
  7. Flinders University, Adelaide, South Australia, Australia
  8. University of Tasmania, Hobart, Tasmania, Australia

Glaucoma is the leading cause of irreversible blindness worldwide. Given its strong genetic component, polygenic risk scores (PRS) offer a promising tool for risk stratification. The latest multi-ancestry glaucoma PRS performs well across major ancestries and is now used clinically, but its utility in underrepresented populations remains unclear. We aimed to calibrate this PRS in seven such groups common in Australia: Vietnamese, Filipino, Sri Lankan, Middle Eastern, Italian, Greek, and Hispanic.

Participants were recruited from Australian hospitals and clinics, supplemented by large research cohorts including the Genetics of Glaucoma study, the Australian and New Zealand Registry of Advanced Glaucoma, the QSkin Sun and Health Study, and CARTaGENE. We inferred genetic ancestry using birth country and ethnic background data from the UK Biobank and 1000 Genomes reference panels. The multi-ancestry glaucoma PRS, developed from more than six million individuals across multiple cohorts, was evaluated using logistic regression to estimate odds ratios (OR) per standard deviation (SD) of PRS, adjusting for age and sex. We compared results with those from well-represented populations (Northwest European, Indian, and Chinese). Population-specific absolute risks were estimated using PRS effect sizes alongside baseline prevalences and age effects from meta-analyses of prior population-based studies.

The PRS was significantly associated with glaucoma risk in all seven target populations. ORs per SD ranged from 1.76 (Filipino) to 3.46 (Sri Lankan). Discriminatory performance was moderate, with AUCs of 0.70–0.79. For most groups, performance did not differ significantly from that of their closest genetically related, over-represented counterparts. Absolute risk varied across ancestries due to differences in PRS effects, baseline prevalences, and age-related risk patterns.

The multi-ancestry glaucoma PRS effectively stratifies risk in diverse, previously underrepresented populations and can inform absolute risk estimation. Further expansion of diverse ancestry cohorts will improve calibration and help ensure equitable access to personalised glaucoma risk prediction.