Abstract
Purpose:
Genome-wide association studies (GWAS) have successfully identified about forty genetic loci associated with refractive error (RE). Despite the relevance of the new findings from a functional perspective, collectively these loci explain a small fraction of RE phenotypic variance or heritability. The aim of this work is to address the question of how much RE heritability can be explained by available platforms and methodologies (direct genotyping, 1000Genomes imputation and direct sequencing).
Methods:
A panel of 1,626 population-based subjects from the TwinsUK cohort was used to calculate the genetic contribution to phenotypic variability. Three different datasets were available for these subjects: directly genotyped, imputation to the 1000Genomes reference panel and whole-genome Next Generation Sequencing (NGS). Contribution to RE phenotypic variability was calculated for each chromosome and over the whole genome using the GCTA package. In addition, the genetic correlation between RE and two other eye quantitative phenotypes, intraocular pressure (IOP) and vertical cup to disk ratio (VCDR) were calculated.
Results:
The portion of phenotypic variability explained by genomic data increases proportionally with the density of the markers: 0.59 for the SNP chip data, 0.67 for the 1000Genomes imputation and 0.74 for the NGS data. A chromosome-by-chromosome comparison confirmed the better performance of the NGS and showed that a number of variants, collectively explaining large portions of heritability, but not represented in the SNP chips used or imputed to the 1000G reference panel may be located in chromosomes 2, 3, 5 and 13 as these were the chromosome for which direct sequencing showed the most improvement. RE also showed strong shared genetic risks with IOP (r=-0.10, SE=0.038, p=0.005) and VCDR (r=0.09, SE=0.04, p=0.006).
Conclusions:
The proportion of phenotypic variance explained by the NGS approached the RE heritability estimated at 0.77 in the same cohort (Lopes et al. 2008). Loci in several chromosomes explain large portions of phenotypic variance but are only captured by direct sequencing. The models also show that many genetic factors explain heritability of more than one eye phenotype. These findings are likely to have implications in future planning and study design of genetic investigation of RE and other ocular phenotypes.
Keywords: 464 clinical (human) or epidemiologic studies: risk factor assessment •
534 gene mapping •
605 myopia