Combinatorial Mismatch Scan (CMS) for loci associated with dementia in the Amish

Population heterogeneity may be a significant confounding factor hampering detection and verification of late onset Alzheimer's disease (LOAD) susceptibility genes. The Amish communities located in Indiana and Ohio are relatively isolated populations that may have increased power to detect disease susceptibility genes.

We recently performed a genome scan of dementia in this population that detected several potential loci. However, analyses of these data are complicated by the highly consanguineous nature of these Amish pedigrees. Therefore we applied the Combinatorial Mismatch Scanning (CMS) method that compares identity by state (IBS) (under the presumption of identity by descent (IBD)) sharing in distantly related individuals from such populations where standard linkage and association analyses are difficult to implement. CMS compares allele sharing between individuals in affected and unaffected groups from founder populations. Comparisons between cases and controls were done using two Fisher's exact tests, one testing for excess in IBS allele frequency and the other testing for excess in IBS genotype frequency for 407 microsatellite markers.

In all, 13 dementia cases and 14 normal controls were identified who were not related at least through the grandparental generation. The examination of allele frequencies identified 24 markers (6%) nominally (p ≤ 0.05) associated with dementia; the most interesting (empiric p ≤ 0.005) markers were D3S1262, D5S211, and D19S1165. The examination of genotype frequencies identified 21 markers (5%) nominally (p ≤ 0.05) associated with dementia; the most significant markers were both located on chromosome 5 (D5S1480 and D5S211). Notably, one of these markers (D5S211) demonstrated differences (empiric p ≤ 0.005) under both tests.

Our results provide the initial groundwork for identifying genes involved in late-onset Alzheimer's disease within the Amish community. Genes identified within this isolated population will likely play a role in a subset of late-onset AD cases across more general populations. Regions highlighted by markers demonstrating suggestive allelic and/or genotypic differences will be the focus of more detailed examination to characterize their involvement in dementia.