A multiethnic genome-wide analysis of 19,420 individuals identifies novel loci associated with axial length and shared genetic influences with refractive error and myopia

Abstract

Introduction: Long axial length (AL) is a risk factor for myopia. Although family studies indicate that AL has an important genetic component with heritability estimates up to 0.94, there have been few reports of AL-associated loci. Methods: Here, we conducted a multiethnic genome-wide association study (GWAS) of AL in 19,420 adults of European, Latino, Asian, and African ancestry from the Genetic Epidemiology Research on Adult Health and Aging (GERA) cohort, with replication in a subset of the Consortium for Refractive Error and Myopia (CREAM) cohorts of European or Asian ancestry. We further examined the effect of the identified loci on the mean spherical equivalent (MSE) within the GERA cohort. We also performed genome-wide genetic correlation analyses to quantify the genetic overlap between AL and MSE or myopia risk in the GERA European ancestry sample. Results: Our multiethnic GWA analysis of AL identified a total of 16 genomic loci, of which 5 are novel. We found that all AL-associated loci were significantly associated with MSE after Bonferroni correction. We also found that AL was genetically correlated with MSE (r_g = -0.83; SE, 0.04; p = 1.95 × 10^-89) and myopia (r_g = 0.80; SE, 0.05; p = 2.84 × 10^-55). Finally, we estimated the array heritability for AL in the GERA European ancestry sample using LD score regression, and found an overall heritability estimate of 0.37 (s.e. = 0.04). Discussion: In this large and multiethnic study, we identified novel loci, associated with AL at a genome-wide significance level, increasing substantially our understanding of the etiology of AL variation. Our results also demonstrate an association between AL-associated loci and MSE and a shared genetic basis between AL and myopia risk.

Keywords: GWAS; axial length; eye biometry; genetics; myopia; refractive errors; single nucleotide polymorphisms (SNPs).

FIGURE 1

Manhattan plot of the multiethnic ancestry GWA meta-analysis of AL in GERA. The y-axis represents the -log₁₀(p-value); all p-values derived from linear regression model are two-sided. The red dotted line represents the threshold of p = 5.0 × 10⁻⁸ which is the commonly accepted threshold of adjustments for multiple comparisons in GWAS. Locus names in blue are for the novel loci and the ones in dark are for the previously reported ones.

FIGURE 2

Correlation of effect sizes across studies (AL vs. myopic refractive error) for the lead 17 AL-associated lead SNPs identified in the current study. Comparison of regression coefficients for AL in 19,420 GERA participants (x-axis) and for MSE in 72,388 GERA participants (y-axis). Pearson’s correlation coefficient = −0.86. Regression Coefficients (betas) are shown, and 95% confidence intervals (CI) are displayed (the horizontal line for each SNP represents the 95% CI of the AL analysis and the vertical line for each SNP represents the 95% CI of the MSE analysis). The 17 AL-associated lead SNPs are shown as black dots and the solid line indicates the line of best fit through the 17 AL lead SNPs.