Abstract
Purpose: :
A genome-wide association study on 600 subjects from the age-related eye disease study (AREDS) was deposited into dbGaP. We sought to determine the extent to which SNPs associated with AMD could be replicated in a second cohort.
Methods: :
SNPs associated with AMD using allelic tests (P <0.0001) reported on dbGaP were identified and characterized by level of support based on Hardy-Weinberg equilibrium (HWE), minor allele frequency (MAF), and the presence of nearby SNPs associated with AMD. The raw genotypes for SNPs across these loci were analyzed with genotype-based association methods that are less susceptible to artifacts than the allelic tests reported in dbGaP. SNPs across the genes in the above loci were selected to tag common blocks of DNA in the population and are being genotyped in a replication cohort of 481 cases with and 311 controls without AMD.
Results: :
A total of 46 previously unreported loci had allelic tests of P<0.0001 in the dbGAP. Of these 46, 20 loci were considered most promising based on a stringent HWE criterion (P>0.01), a MAF > 0.01, and one or more adjacent SNPs associated with AMD as reported on dbGaP. Genotypic association tests (Fisher’s exact test and logistic regression with log-additive genetic model) suggested that 14 (70%) of these 20 loci had one or more SNPs that were associated with AMD (P<0.0001). Genotyping of the replication cohort on these 14 loci, and other loci with genotypic association, is underway and will be analyzed.
Conclusions: :
In addition to identifying previously established loci (CFH, ARMS2, and BF/C2) and a novel variant (C3) replicated recently, at least 14 additional novel loci may be associated with AMD based on robust genotypic association analysis. Allelic tests are dependent upon the assumption of HWE and slight deviation from HWE can result in biased tests, unlike genotypic tests for association. We found that only 70% of the allelic tests of the most promising loci were supported by genotypic analysis. Use of allelic test for genetic association studies (e.g., dbGaP) should be avoided because genotypic tests more accurately reflect the underlying genetic association. The results of a replication study of loci with genotypic assocation tests will be presented.
Keywords: age-related macular degeneration • genetics