Detecting genetic loci for preservation of cognition in the Midwestern United States Amish

Abstract Background Alzheimer’s Disease (AD) is the most common form of dementia and has limited treatments. While AD risk has a large genetic component, under 50% of the genetic risk has been identified. By focusing on identifying genetic variants that delay or prevent the onset of AD, our goal is to identify new associated genes and variants. To detect such loci, we studied cognitive preservation in the Midwestern Amish, a genetically and culturally isolated population of European descent. Methods We studied 946 Amish individuals (ages 76‐95) from Ohio and Indiana. Participants were classified as cognitively impaired (33.8%) or cognitively unimpaired (66.2%) by consensus review of cognitive examinations. These individuals were all connected in a 15‐generation 8,222‐person pedigree. This pedigree was divided into 104 sub‐pedigrees via PedCut for linkage analysis. MERLIN was used for autosomal linkage analysis, while MINX was used for X‐chromosome linkage analysis. The association analysis, corrected for genetic relationship, used GENESIS for autosomes and XWAS for the X chromosome. Results Over 250,000 SNPs were used for association and two‐point linkage analyses, and a subset of 5,294 uncorrelated SNPs was used for multipoint linkage analyses. on 15 different chromosomes, with the highest two‐point heterogeneity LOD score (HLOD = 5.85) on chromosome 2 (∼79Mb) and highest multipoint HLOD (3.55) on chromosome 12 (∼25Mb), neither overlapping with known AD genes. Two linkage results overlapped with known AD genes (Chr 6: CD2AP; Chr 11: PICALM) and do not overlap with association results. Significant (p ≤ 6.4x10 ‐7 ) and suggestive (p ≤ 1x10 ‐4 ) thresholds for association were determined using SimpleM. While no significant associations were found, suggestive associations were found for 103 SNPs (11 loci) across 10 chromosomes. Three of these (chr 11: MS4A2; chr 14: SLC24A4; chr 16: IQCK) fall near known AD genes. These significant and suggestive regions are being followed up currently with fine mapping. Conclusion Preliminary analyses suggest that using cognitive preservation as our phenotype of interest identifies both known AD loci and novel regions that warrant further evaluation and may lead to a further understanding of both AD and cognitive preservation.