Genome-wide risk prediction of primary open-angle glaucoma across multiple ancestries

BACKGROUND: Primary open-angle glaucoma (POAG) is a leading cause of blindness worldwide. The disease is often only diagnosed after retinal ganglion cell damage has occurred, with current treatments unable to restore lost vision. Developing risk identification tools for POAG will help enable timely diagnosis and prevent irreparable damage from occurring, especially for ancestry groups (such as African (AFR)) where the disease prevalence is high. Given the heritable nature of POAG, we aim to develop a polygenic score (PGS), which could facilitate earlier POAG risk detection for timely prevention and diagnosis. METHODS: We applied a multi-ancestry multi-trait approach to build powerful PGS for POAG. We first integrated the new and existing genetics data on POAG and two key endophenotypes, intraocular pressure (IOP) and vertical cup-to-disc ratio (VCDR). We then leveraged the shared POAG genetic information across European (EUR), AFR and Asian ancestries and between POAG and each of IOP and VCDR to develop PGS for POAG risk prediction. We systematically assessed the PGS prediction power and risk stratification ability in POAG cohorts of different ancestries. RESULTS: Our newly developed PGS showed improved accuracy compared to previous PGS for POAG risk prediction in EUR ancestry. We showed the transferability of PGS based on EUR ancestry in the prediction of POAG status in AFR and Asian ancestries. Utilizing the shared genetic information across ancestries further improved PGS prediction power for POAG in AFR and East Asian (EAS). For individuals with South Asian ancestry, those in the top PGS decile were diagnosed ~18 years earlier than those in the bottom decile. For AFR ancestry, individuals in the top percentile had an odds ratio of 4.08 (95% CI: 2.33-7.45) compared with the remainder of the population using the newly developed AFR-specific PGS. CONCLUSIONS: In the current study, we developed PGS for POAG risk prediction in EUR, Asian and AFR populations. These PGS, to our best knowledge, are the most powerful PGS currently for POAG risk screening and stratification. We believe that our study will lead to improved POAG detection across diverse populations in the future by enabling targeting clinical screening of people at high levels of genetic risk.