Abstract
Purpose :
Geographic atrophy (GA) and drusen are important features in non-neovascular age-related macular degeneration (AMD). Detailed GA and drusen measurements could be promising biomarkers for staging and monitoring of AMD. However, no genome-wide search for the genetic risk underlying these measures, either shared with AMD or independent, has been done. Here, we conduct a genome-wide association study (GWAS) on these measures in Amish from Ohio, Pennsylvania, and Indiana to identify associated variants.
Methods :
Using an Illumina MEGAex with custom content, we performed exome chip genotyping and analyzed common variants with at least 1% allele frequency after extensive quality control. GA and drusen measures were obtained from retinal optical coherence tomography (OCT) volume scans. A total of 1,136 Amish adults from Ohio, Pennsylvania, and Indiana with a family history of AMD were included in the analyses. Three traits were measured including GA area along with both drusen elevation area and volume. A linear mixed model with genetic similarity matrix, accounting for relatedness, was used to test for association between these traits and exome chip data.
Results :
Across the three traits, we identified three loci with genome-wide significance (p<5.0x10-8) for quantity of drusen or GA among those with any presence of drusen or GA. Each of these are in or downstream of known genes (GLRX3: p=1.6x10-9, MAST4: p=3.5x10-8, MIR9-3: p=1.8x10-8). Additionally, we found five loci with suggestive (p<5.0x10-6) associations that were in or downstream of known genes (CRYL1: p=5.8x10-8, KIF2B: p=4.6x10-8, ZNF609: p=5.8x10-7, PLD3: p=9.7x10-7, CTNND2: p=7.0x10-7) and one locus that did not occur near known genes on chromosome 5 (p=7.6x10-7). We also identified one locus that is suggestive for any presence of drusen or GA occurring in a known gene (LHPP: p=6.6x10-7) and one locus not near known genes on chromosome 1 (p=9.2x10-8). Quantile-quantile plots for each of the traits did not suggest evidence of stratification, technical bias, or errors.
Conclusions :
We identified two loci that are potentially associated with presence of any drusen or GA and eight loci that may be associated with quantity of drusen or GA. The strongest of these variants were located on chromosomes 5, 10, and 15. These findings will help advance understanding of the biological mechanism of AMD development and improve understanding of the etiology of the disease.
This is a 2020 ARVO Annual Meeting abstract.