Deep-learning-derived glaucoma-related endophenotypes enable novel genome-wide genetic and functional discovery

The genetic architecture of primary open-angle glaucoma (POAG) remains largely unexplained due to the reliance of previous genome-wide association studies (GWAS) on imprecise phenotypes from electronic health records. We overcome this with a disease-trained, task-transfer machine learning (ML) framework that learns glaucoma-related damage patterns from a clinical repository of 18,985 optical coherence tomography (OCT) scans from 8323 glaucoma patients. By applying the derived endophenotypes to 39,146 UK Biobank participants, we performed GWAS in European, African, and Asian ancestral groups, followed by cross-ancestry meta-analyses. We identified 36 and 43 LD-independent GWAS loci that passed genome-wide significance in the EUR and cross-ancestry meta-analyses, respectively. About two-thirds of the identified loci overlapped with previously reported POAG-related associations, demonstrating the validity of our approach. Importantly, more than a third of the loci were novel to glaucoma. Extensive functional analyses, including Bayesian colocalization analysis, gene-based association tests, Mendelian randomization, and single-cell enrichment analysis, converged on 11 putative causal genes, five of which were novel to glaucoma. These genes implicate Wnt-mediated outflow dysfunction and retinal ganglion cell vulnerability as candidate mechanisms in POAG. Our findings expanded POAG genetic associations, provided mechanistic insights at cell-type resolution, and proposed plausible putative causal genes.