Premium
O3‐05‐04: Admixture analysis of Alzheimer's disease in caribbean hispanics
Author(s) -
Tosto Giuseppe,
Lee Joseph H.,
Vardarajan Badri N.,
Cheng Rong,
Reyes-Dumeyer Dolly,
Barral Sandra,
Reitz Christiane,
Mayeux Richard
Publication year - 2015
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2015.07.264
Subject(s) - genetic admixture , confounding , demography , disease , ethnic group , allele , genetic association , gerontology , biology , medicine , genetics , gene , genotype , population , single nucleotide polymorphism , sociology , anthropology
throughout the genome with LOAD risk in 4,831 LOAD cases and 3,174 cognitively normal controls in the Alzheimer’s Disease Genetics Consortium (ADGC). Methods:We called CNVs using quality-controlled, high-density GWAS genotyping on the Illumina 660w and HumanOmniExpress SNP arrays in 8,005 participants from waves 1-6 of the ADC (Alzheimer’s Disease Center) subsets of the ADGC using the PennCNV program. Quality control excluded variants from telomeric, centromeric, and HLA regions and samples with excess standard deviation of allelic intensity [SD(LRR)>0.3] or G/C base content waviness factor [jGCWFj>0.05 or more than 100 CNVs genome-wide. We analyzed 31,007 duplications (n1⁄47,411) and 63,200 deletions (n1⁄47,628) using joint SNV-CNVassociation in ParseCNV. Results: We observed an average of 12.16 CNVs per individuals (4.01 duplications and 8.16 deletions), and no significant quantitative difference between cases and controls. Joint association testing in ParseCNV identified 3 significant deletions enriched in cases, and 1 deletion and 1 duplication enriched in controls (OR>2; contained 10 more SNPs; uncorrected two-tailed P<5310), all with frequencies<3%. Joint SNV/CNV gene-based testing identified 17 genes in or near deletions and 7 genes in or near duplications associated with LOAD at gene-wide significance (uncorrected two-tailed P<5310). We estimated heterogeneity in associations by genotyping platforms using METAL, which identified CNVassociations of P<5310 in 5 genes in or near deletions (EYS, TYRP1, RAB9BP1, ASCC3, FSTL5) but not duplications. Conclusions: Our findings suggest several rare deletion CNVs with strong effects may contribute to LOAD risk, however confirmation in independent datasets is needed. Additional work to clarify the frequency of these CNVs in large population samples is ongoing using case-only rare CNV annotation in population databases.