Purchase this article with an account.
Xiaohui Li, Yelena Bykhovskaya, Talin Haritunians, David Siscovick, Anthony J. Aldave, Loretta B. Szczotka-Flynn, Sudha K. Iyengar, Jerome I. Rotter, Kent D. Taylor, Yaron S. Rabinowitz; Several Susceptible Regions Are Associated With Keratoconus Using Genome-wide Association And Confirmation Panels. Invest. Ophthalmol. Vis. Sci. 2011;52(14):4385.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To identify genetic susceptibilities of keratoconus (KC) using a genome-wide association study (GWAS) and a replication design.
We first performed a GWAS using 222 KC Caucasian patients and 3324 controls and identified several KC associated regions (p = 10-6). Genotyping of GWAS was performed using Illumina 370K bead chips. After excluding SNPs with minor allele frequencies of <0.05 and/or with a call rate of < 0.9, ~320K SNPs were analyzed. Principle components analysis was performed to test population stratification between cases and controls. Further fine mapping of selected regions with 4650 SNPs were genotyped using an independent case-control panel (304 cases and 518 controls) and a family panel (312 subjects in 70 families) to confirm the original GWAS findings. Association tests and logistic regression models implemented in PLINK (v1.07) were used for case-control samples and transmission disequilibrium test (TDT) implemented in GeneHunter 2.1 was used for the family panel.
Using the original GWAS, we identified three regions (3p26, 2q21.3 and 19q13.4) with 3 SNPs reaching genome-wide significance (p<5×10-7). The top two SNPs, rs6442925 at 3p26 and rs4954218 at 2q21.3, are associated with KC with p values of 6.5×10-8 and 2.4×10-7. Several genes are clustered in the 2q21.3 region, including the RAB3GAP1, YSK4 and ZRANB3 genes. Seven other regions with significance levels of less than 5×10-6 were also identified using the original GWAS. Fine mapping of independent case-control samples confirmed the top 2 regions (3p26: p~0.01 to 0.05; 2q21.3: p~4.0×10-5 to 0.05) and one region at 7q31 (original GWAS p of 2.4×10-6, confirmation p~0.02 to 0.04). The most significant SNP from meta-analysis of same SNPs by combining original GWAS and confirmation panel was rs4954218 at 2q21.3 (p=2.8×10-8). Interestingly, the family panel also identified the same trends of associations in 2q21.3, although they were not statistically significant.
Using GWAS and replication panels, we identified 3 susceptibility regions associated with KC, the most significant being 2q21.3, which contains a cluster of genes that may be associated with eye disease.
This PDF is available to Subscribers Only