Premium
S1‐01‐01: Finding new susceptibility genes for Alzheimer's disease: The GERAD consortium
Author(s) -
Gerrish Amy,
Williams Julie
Publication year - 2011
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.2011.05.198
Subject(s) - genome wide association study , snp , single nucleotide polymorphism , genetics , genetic association , biology , snp array , allele , allele frequency , gene , genotype
Background: In 2009 we published a GWAS study of Alzheimer’s disease (AD) in which we identified two new genome-wide significant susceptibility loci: CLU and PICALM. We also observed more variants with P < 1~ A— 10 5 than expected by chance. One of these, BIN1, became genome-wide significant when our data were added to that of the CHARGE and EADI consortia. To identify new common susceptibility variants for AD, we undertook a powerful staged association study comprising 19,870 cases and 39,846 controls. This identified two new susceptibility variants in rs3764650 at ABCA7 (meta-P1⁄44.5~ A—10 ) and rs610932 at the MS4A gene cluster (meta-P 1⁄4 1.8~ A—10 ). We further tested four loci, which showed suggestive evidence for association in the ADGC GWAS. Three became genome-wide significant following analyses in GERAD + samples: rs9349407 at CD2AP (meta-P 1⁄4 8.6~ A—10 ), rs3865444 at CD33 (metaP 1⁄4 1.6~ A—10 ) and rs11767557 at EPHA1 (meta-P 1⁄4 6.0~ A—10 ). We will present new analyses of imputed GWAS data, association with CNVs, homozygosity mapping and mitochondrial variants. Methods: Using IMPUTE, untyped SNPs were imputed using the latest 1000 genomes reference panel. CNVs were identified with PennCNV using SNP intensity signals extracted from Beadstudio. These were then analysed using PLINK. Runs of Homozygosity (>1Mb) were tested for excess burden in cases compared to controls using PLINK. Mitochondrial allele frequencies were compared in cases and controls either on an individual SNP by SNP basis using Fisher’s exact (P) or across the entire data set by permuting the disease status. Results: Ten genes have now been established as risk loci for AD. Within these, certain functional processes have emerged which implicate changes in the immune response and inflammation processes, lipid processing and endocytosis within AD. CNVanalyses also point to dysfunction in these processes as well as novel pathways. Homozygosity mapping and the association of mitochondrial variants will also be presented. Conclusions: We have made significant progress in defining the genetic architecture of AD. However, more genes remain to be identified. We are currently contributing to further powerful analyses (I-GAP) and are testing for relationships with AD symptoms, rate of decline and for associations that cross diagnostic boundaries by combining our data with data from other neurodegenerative diseases and disorders associated with AD.