Genome-Wide Association Studies

Although karyotyping and linkage analysis can still be used to identify disease-associated genes, most research is now centered on GWAS and NGS. The International HapMap Project, which followed the creation of the human gene map, compared the DNA sequence of 1184 reference individuals from 11 global populations (creating a catalog called the HapMap) to identify regions of variation between individuals and racial groups. By using the HapMap to study individuals from a similar population, genetics researchers find that many people will share a series of SNPs or a haplotype. Thus, it is possible to test only one or a few SNPs but to infer a large number of adjacent SNPs by imputation. Chip or bead platforms enable the investigation of 100,000 to millions of SNPs across the genome, forming the basis of a GWAS.

Figure 5-4 Schematic representation of the original Sanger dideoxy chain-termination sequencing method. The results produced are shown in step 3: DNA fragments resolved on a polyacrylamide gel and a sequencing trace from a modern automated sequencing machine.

(Original figure from Oxbridge Biotech Roundtable; redrawn by Mark Miller.)

Results of a GWAS are usually presented in a Manhattan plot, so named because it brings to mind the New York City skyline. In a Manhattan plot, the chromosomes are arranged in order along the x-axis, and the P value (as −log P) of the association of the disease or trait with the particular SNP at that chromosomal location is given on the yaxis. Figure 5-6 shows a Manhattan plot for glaucoma. A significant gene association (threshold ≈5 × 10^–8) will often have multiple adjacent SNPs at high levels of significance, and thus a column of points will rise on the plot. It is rare that the SNPs themselves are the disease-causing mutations. Usually they are linked in the haplotype to the mutation, which is why researchers will then use fine-mapping of the region by looking at a large number of SNPs in the nearby region.

Figure 5-5 Schematic summary of the wholege-nome shotgun sequencing method of next-generation sequencing (NGS). At a basic level, all NGS technologies use the same princi ple: fragment the DNA, add primers/adapters, amplify, and sequence. In whole-genome shotgun sequencing, a DNA sample is randomly broken into numerous small fragments that are then sequenced using the chain-termination method. Multiple overlapping DNA fragments produced from numerous repetitions of this process are then assembled into a single continuous sequence on a computer program.

(Original figure from Oxbridge Biotech Roundtable; redrawn by Mark Miller.)

Combining numerous studies, usually of multiple ethnic groups, in meta-analyses allows for identification of additional associated gene regions. Figure 5-7 shows how GWAS meta-analyses combine data from individual GWAS. Figure 5-8 shows the meta-analysis of GWAS for age-related macular degeneration (AMD), with 19 loci now identified. The effect size for all of these genes is usually small, but cumulatively they account for approximately 40% of AMD heritability.

Figure 5-6 Manhattan plot for glaucoma, identifying the CDKN2BAS region at 9p21 and the SIX1/SIX6 region at 14q23.

(Reproduced with permission from Wiggs JL, Yaspan BL, Hauser MA, et al. Common variants at 9p21 and 8q22 are associated with increased susceptibility to optic nerve degeneration in glaucoma. PLoS Genet. 2012;8(4):e1002654.)

Figure 5-7 Meta-analysis of Manhattan plots.

(Reproduced with permission from Manolio TA. Genomewide association studies and assessment of the risk of disease. N Engl J Med. 2010;363(2):166–176.)

Figure 5-8 Manhattan plot for age-related macular degeneration meta-analysis identifying numerous associated genes.

(Reproduced with permission from Fritsche LG, Chen W, Schu M, et al. Seven new loci associated with age-related macular degeneration. Nat Genet. 2013;45(4):433–439.)

Comparison of GWAS from different ethnic groups can help clarify whether the SNP itself is disease causing or just linked to the true disease-causing mutation(s). An example is the LOXL1 gene, which is associated with pseudoexfoliation. One SNP was associated with disease in the Caucasian population, but disease was associated with the alternate SNP in the Japanese population (Fig 5-9A). Thus, it is unlikely that this SNP is actually disease causing but more likely that different SNPs are associated with the true disease-causing mutation in East Asian and white, or Caucasian-derived, populations. In contrast, another SNP had equivalent association in both populations (Fig 5-9B).

For a catalog of GWAS including ophthalmic studies, see the European Molecular Biology Laboratory–European Bioinformatics Institute (EMBL-EBI) catalog at http://www.ebi.ac.uk/gwas/.

Excerpted from BCSC 2020-2021 series: Section 2 - Fundamentals and Principles of Ophthalmology. For more information and to purchase the entire series, please visit https://www.aao.org/bcsc.