APOE is not associated with Alzheimer disease: a cautionary tale of genotype imputation

Abstract

With the advent of publicly available genome-wide genotyping data, the use of genotype imputation methods is becoming increasingly common. These methods are of particular use in joint analyses, where data from different genotyping platforms are imputed to a reference set and combined in a single analysis. We show here that such an analysis can miss strong genetic association signals, such as that of the apolipoprotein-e gene in late-onset Alzheimer disease. This can occur in regions of weak to moderate LD; unobserved SNPs are not imputed with confidence so there is no consensus SNP set on which to perform association tests. Both IMPUTE and Mach software are tested, with similar results. Additionally, we show that a meta-analysis that properly accounts for the genotype uncertainty can recover association signals that were lost under a joint analysis. This shows that joint analyses of imputed genotypes, particularly failure to replicate strong signals, should be considered critically and examined on a case-by-case basis.

Figure 1. Plot of –log10(p-value) and missing data in the region surrounding apoE

The plot shows results of the imputation analysis in our GWAS (Fig 1A), the Reiman et al GWAS (Fig 1B), the joint analysis of both GWAS before quality-control (Fig 1C), and the joint analysis after quality-control (Fig 1D). The points refer to –log10 (p-value) of the trend test. Open circles are SNPs that were genotyped on the platform for 1A and 1B. Closed circles mark SNPs inferred by the imputation algorithm. SNPs with crosses were monomorphic and not tested. The bar graph represents the percent of missing data of the tested SNPs.

Figure 2. Patterns of linkage disequilibrium (LD) around apoE

Plot of LD around the apoE gene. LD is measured by r², in the Haploview program, and is calculated from the HapMap CEU parents. The table shows p-value of the trend test after imputation, and percent of samples with missing genotypes for that SNP. SNPs that were genotyped in the initial studies are outlined with black boxes.