Abstract
Purpose :
Refractive development is known to be mediated by both genetic and environmental factors. More than two-dozen loci for refractive phenotypes, discovered via genomewide association studies (GWAS), have been reported. Environmental factors and their measurable surrogates such as educational attainment, are also strongly associated with refractive error. We investigated the main effects and interaction effects of genetic markers and education in a large GWAS meta-analysis.
Methods :
We performed GWAS in 40,036 adults from 25 studies of European ancestry, and 10,315 adults from 9 studies of Asian ancestry and performed a joint meta-analysis to test main SNP and SNPxEducation interaction effects on refractive error. Participant ages ranged from 20 to 94 years; the mean age in individual studies ranged from 25 to 79. Education was dichotomized in all cohorts based on the number of years of formal education (<12 years vs >= 12 years). Refractive error was coded as the mean spherical equivalent refraction. Markers were annotated to genes if they were within 10 kb the gene. Multi-marker gene-based analyses were performed on summary p-values using a multiple linear principal components regression approach. Empirical p-values for 17,800 genes were obtained via permutation. A Bonferroni-corrected p<3x10e-6 (~0.05/17,800) was considered statistically significant.
Results :
In the European ancestry cohorts, 22 genes attained genomewide significance. No genes attained the significance threshold in the Asian cohorts. When both cohorts were combined (n=50,351), 27 genes were significantly associated with spherical equivalent refraction. Of these, 13 genes in 7 different loci appear to be novel. Genes in other well-replicated loci for refraction (e.g., GJD2 & LAMA2 loci) were not significantly associated with refraction in our gene-based analyses, likely because the association signals were intergenic.
Conclusions :
Using gene-based analyses of gene-environment-wide interaction studies, we identified 27 genes that were significantly associated with spherical equivalent ocular refraction, including 13 genes in 7 loci that appear to be novel. Other loci known to be associated with refractive phenotypes did not show statistical significance at the single-gene level, suggesting that intergenic regulatory elements may be implicated in refractive error at these loci.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.