Multi-ancestry genome-wide association analyses of refractive error augment genetic discovery and polygenic prediction

Refractive errors (REs) affect over half of the global population, with consequences ranging from blurred vision to blindness. Here we conducted ancestry-stratified and cross-ancestry meta-analyses of genome-wide association studies for RE in people of European (n = 1,495,159), East Asian (n = 121,172) and African (n = 144,737) ancestries. The cross-ancestry meta-analysis identified 932 RE-associated variants, including 241 previously unknown associations, four East Asian-specific associations and one African-specific association. Statistical fine-mapping pinpointed 16 high-confidence putative causal variants, and gene prioritization analyses highlighted 23 genes involved in eye development. We constructed an enhanced polygenic predictor incorporating functional annotations that explained 21.4% of RE variation, effectively stratified the onset, progression and severity of myopia, and achieved an area under the receiver operating characteristic curve of 0.806 for predicting high myopia. Our multi-ancestry genome-wide association study expands substantially the catalog of genetic variants for RE and demonstrates the potential clinical utility of polygenic prediction in identifying high-risk people across diverse populations.