**Background:** Primary open-angle glaucoma (POAG) is a leading cause of irreversible blindness worldwide. It is often asymptomatic until later stages, causing optic nerve damage manifested by cupping and visual field loss. Large vertical cup-to-disc ratio (VCDR) and elevated intraocular pressure (IOP) are two key POAG endophenotypes. Previous genome-wide association studies (GWASs) identified 127 loci, collectively explaining 9.4% of the familial risk. However, many risk loci remain undiscovered and their biological functions are largely unknown. Multitrait methods have demonstrated substantial improvements in power for uncovering novel genetic loci when incorporating data from related endophenotypes. Both VCDR and IOP are highly genetically correlated with glaucoma (genetic correlation 0.50 [s.e.m. = 0.05] and 0.71 [s.e.m. = 0.04], respectively). This study leveraged new and existing genetic data for POAG, VCDR, and IOP to perform a large-scale multitrait analysis of GWAS (MTAG) to identify novel POAG loci, integrate data across ancestries, and utilize omics datasets to improve understanding of biological mechanisms and druggable target discovery.
**Methods:** The study first performed an MTAG in the European ancestry population, including GWASs for POAG (29,241 cases and 350,181 controls) and its two key endophenotypes, VCDR (n = 111,724) and IOP (n = 153,604). The identified novel POAG loci from MTAG were replicated in a large-scale independent glaucoma GWAS (23andMe, Inc. study, 84,910 cases and 2,736,075 controls). A multiancestry meta-analysis for POAG was then conducted, combining the MTAG POAG output from the European ancestry population and samples from Asian (6,935 cases and 39,588 controls) and African (3,281 cases and 2,791 controls) ancestry populations. Fine-mapping and post-GWAS functional analytical approaches were applied to prioritize genetic findings, identify druggable targets, and characterize potential biological mechanisms. These included gene-based and pathway analyses, classification of POAG loci into VCDR- or IOP-specific SNPs, colocalization analysis with eQTL/sQTL data, transcriptome-wide association study (TWAS), proteome-wide association study (PWAS), and Mendelian randomization (MR) to map the effects of plasma proteome on POAG risk. Drug target prioritization was performed using multiple lines of genetic evidence.
**Key Results:** In the MTAG analysis of European ancestry, 263 independent loci for POAG were identified, of which 81 were novel (not within ±500 kb of previously known loci). The proportion of familial risk explained by genome-wide significant independent SNPs was 14.1%, a 50% increase over the previous estimate of 9.4%. The 81 completely novel loci contributed 2.5%, with the remainder of the difference (2.2%) attributable to additional independent SNPs within previously reported loci. Replication in the 23andMe study showed that 60% of SNPs (n = 156) passed genome-wide significance (P < 5 × 10^-8), 85% (n = 223) were significant after Bonferroni correction (P < 0.00019), and 98% (n = 256) reached nominal significance (P < 0.05). The effect sizes showed very high concordance (Pearson's coefficient 0.97, P = 5.99 × 10^-154). The multiancestry meta-analysis identified 312 independent loci, including 109 novel loci. Replication in 23andMe showed 56% (n = 169) passed genome-wide significance, 79% (n = 240) passed Bonferroni correction, and 98% (n = 296) reached nominal significance, with high concordance (Pearson's coefficient 0.96, P = 1.22 × 10^-164). Gene-based analysis identified 355 significant genes, and pathway analysis uncovered 32 pathways, including those involved in collagen formation, blood vessel development, and cardiovascular system development. Classification of POAG loci into VCDR- or IOP-specific SNPs identified 92 VCDR-specific and 171 IOP-specific SNPs. Colocalization analysis with eQTL/sQTL data prioritized causal genes for 139 (52.9%) of the replicated MTAG POAG loci. TWAS identified 86 genes associated with POAG risk. MR analysis identified 33 proteins potentially causally associated with POAG risk. Drug target prioritization identified 69 potential drug target genes, with 17 having at least two levels of genetic evidence. Genetic correlation and MR analyses identified 24 traits genetically correlated with POAG, VCDR, or IOP, and 14 traits showing putatively causal effects on POAG risk, including multiple sclerosis, systemic lupus erythematosus, type 2 diabetes, and immune cells.
**Clinical Implications:** This study nearly triples the number of known POAG risk loci, explaining a substantially larger proportion of familial risk. The identification of 69 potential drug targets, including 17 with strong genetic support, provides a rich resource for developing new therapies. Importantly, the classification of loci into VCDR- and IOP-specific effects highlights potential neuroprotective targets that act independently of IOP, addressing a major unmet need in glaucoma treatment. The associations with immune disorders suggest a role for immune dysregulation in glaucoma pathogenesis, opening new avenues for therapeutic intervention. These findings enable new drug development for this common cause of blindness and provide insights into the underlying biological mechanisms.