Genetic profile for five common variants associated with age-related macular degeneration in densely affected families: a novel analytic approach

Abstract

About 40% of the genetic variance of age-related macular degeneration (AMD) can be explained by a common variation at five common single-nucleotide polymorphisms (SNPs). We evaluated the degree to which these known variants explain the clustering of AMD in a group of densely affected families. We sought to determine whether the actual number of risk alleles at the five variants in densely affected families matched the expected number. Using data from 322 families with AMD, we used a simulation strategy to generate comparison groups of families and determined whether their genetic profile at the known AMD risk loci differed from the observed genetic profile, given the density of disease observed. Overall, the genotypic loads for the five SNPs in the families did not deviate significantly from the genotypic loads predicted by the simulation. However, for a subset of densely affected families, the mean genotypic load in the families was significantly lower than the expected load determined from the simulation. Given that these densely affected families may harbor rare, more penetrant variants for AMD, linkage analyses and resequencing targeting these families may be an effective approach to finding additional implicated genes.

Figure 1

Estimated prevalence of advanced AMD cases as a function of genetic load. Using genotyping data from a case–control study of AMD, 36 risk categories were determined on the basis of genotypes at the five common risk loci. The estimated risk of developing advanced AMD for each of these risk categories was calculated. These risk categories are sorted by the estimated risk of advanced AMD. The table below the figure shows the specific genotypes for each bin. For CFH, the L/M/H designations correspond to low-, medium- and high-risk haplotypes. The LOC387715 L/H designation corresponds to low- and high-risk alleles. For C2/BF, the L or H designation corresponds to low- or high-risk genotypes.

Figure 2

Distribution of mean genotypic load among families. We calculated the mean familial genotypic load for individuals in the 65 families who have genotype and phenotype information available on all siblings and in the 257 families who did not have genotype and/or phenotype information available on all siblings. The number of individuals in each genotypic load category is shown here for the 65 (a) and 257 (b) families. There was no statistically significant difference between these two distributions (P=0.11, χ²-test).

Figure 3

Observed vs expected genotypic load among families. The mean genotypic load for each family was plotted against the mean genotypic load for a family of the same configuration, as determined by simulation. Each dot represents one family. Triangles mark the lower 2.5% confidence bound, whereas diamonds mark the upper 97.5% confidence bound. The slope of the line for the individual families' genotypic loads is 0.4594.

Figure 4

(a) Distribution of slopes for plots of mean genotypic load across all simulation families vs individual families' genotypic loads from each simulation subset of 322 families, by family configuration. The arrow indicates the location of the slope from Figure 3 within the tail of this distribution. (b) Distribution of slopes for plots of mean genotypic load across all simulation families, in which the simulation included a rare, high-penetrance allele, vs genotypic loads from each simulation subset of 322 families, by family configuration. Please note the difference in the x axis scale between panels a and b. The distribution of values in panel b is shifted to values lower than that of panel a.