Rare variants contributing to age-related macular degeneration - Results from the International AMD Genomics Consortium. W. M. Igl for the International AMD Genomics Consortium Department of Genetic Epidemiology, University of Regensburg, 93053 Regensburg, Bavaria, Germany.
Purpose: Age-related macular degeneration (AMD) is a common cause of blindness in older people with strong genetic contribution from common variants (CVs). In contrast, the role of rare variants (minor allele frequency1%) to AMD has not yet been studied genome-wide. To delineate the role of rare variants (RVs) in AMD, we genotyped about 50,000 samples at a single genotyping center with a custom-modified Illumina exome chip, specifically based on sequencing experiments.
Methods: We analyzed currently available genotype data of 23,362 unrelated individuals of European descent, totaling 12,030 AMD cases vs. 11,332 controls of comparable age. We tested 103,947 autosomal, bi-allelic RVs for association with AMD using Fishers Exact Test. All results were genomic controlled and tested for significance at =0.05/100,000=5E-7. Sensitivity analyses using Exact and Firths bias-corrected logistic regression adjusted for sex, age, ancestry principal components, and/or known CVs (Fritsche et al., NatGenet 2013) were performed for significant variants.
Results: Among the significant RVs were the known functional variant rs121913059 (R1210C, CFH, Effect Allele Frequency (T)=0.00013, odds ratio=24.88, p=2.06E-19) and the newly described rs147859257 (K155Q, C3, EAF(G)=0.0040, OR=3.12, p=1.21E-24). Additional novel RVs were discovered at loci known for common AMD variants: CFH/CFHlike (15 variants, 6.51E-31p4.83E-8), CFB/C2 (2 variants, 1.27E-7p3.27E-7), and ARMS2/HTRA1 (2 variants, 3.71E-11p2.91E-10). One locus as yet unpublished for AMD was detected with a sole significant RV in the C9 gene (p=5.86E-16), which is part of the complement system. The effects were not substantially modified by adjusting for sex, age, or principal components. Several of the signals were substantially weakened by conditioning on known CVs; the known variant in CFH (rs121913059, p=2.41E-12), the C3 variant (rs147859257, p=9.28E-19), and six other novel variants in CFH/CFHlike (2.10E-16p4.70E-7) and C9 (p=1.64E-11) remained significant.
Conclusions: Using world-wide sample collections, this rare variant genome-wide association study (rGWAS) has identified strong associations with AMD in known and novel loci. While some RV effects are independent from CVs highlighting novel AMD mechanisms, others coinciding with CV signals potentially depict underlying functional entities worth further follow-up.
You may contact the first author (during and after the meeting) at