Meta-Analysis of Genome-Wide Association Studies in Myopia in Nine Populations.

Myopia is a common refractive error which affects at least a third of most populations. Both genetic and environmental factors influence myopic development. It has a significant impact on the lives of affected individuals and carries high economic costs associated with treatment, loss of productivity and co-morbidity from vision impairment. Recent genome-wide association studies (GWAS) have identified a number of loci associated with myopia and refractive error. Here we report results of a large meta-analysis of myopia in nine cohorts, for a total of 17,787 individuals of European ancestry and replication in a further 8 cohorts for a total of 7953 individuals. Genotypes in each population were imputed to HapMap2 and analyzed separately by each group. Cases were defined as a spherical equivalent of -1 diopters (D) or worse and controls were defined as > 0D. Individuals between 0 and -1D were coded as unknown. Analyses were performed including age, sex and years of education, plus the first 3 principal components to adjust for population structure. Meta-analysis was performed in METAL using the sample size schema. Due to large differences in numbers of cases and controls for some studies, effective sample sizes were calculated using the formula recommended by the authors of METAL. Genomic control was used to adjust for any residual structural differences between populations. 3 SNPS were identified as genome-wide significant with P < 5e-8, rs10113215 on chromosome 8q12.1 and rs1370156 and rs2028099 both in a previously reported locus on chromosome 15q14. For replication, SNPs with P < 1e-5 were identified and all SNPs within 500kb each side of that SNP selected, for a total of 20,431 SNPs. The replication threshold was set by calculating the effective degrees of freedom using the Ramos method. SNPs will be considered to replicate where the p value < 0.0026. The replication analyses are ongoing and will be presented.