Exome array analysis of 30,582 individuals confirms late-onset Alzheimers disease (LOAD) risk from common variants and identifies novel rare LOAD susceptibility variants: the International Genomics of Alzheimers Project (IGAP). A. C. Naj1, S. J. van der Lee2, M. Vronskaya3, R. Sims3, J. Jakobsdottir4, C. van Duijn2, L.-S. Wang5, P. Amouyel6, S. Seshadri7, J. Williams3, G. Schellenberg5, International Genomics of Alzheimer's Project (IGAP) 1) Biostatistics & Epidemiology, University of Pennsylvania Perelman School of Medicine, Philadelphia, PA, USA; 2) Department of Epidemiology, Erasmus MC University Medical Center, Rotterdam, Netherlands; 3) Institute of Psychological Medicine and Clinical Neurosciences, Medical Research Council (MRC) Centre for Neuropsychiatric Genetics & Genomics, Cardiff University, Cardiff, Wales, UK; 4) Icelandic Heart Association, Kopavogur, Iceland; 5) Department of Pathology and Laboratory Medicine, University of Pennsylvania Perelman School of Medicine, Philadelphia, PA, USA; 6) Institut Pasteur de Lille, Lille, France; 7) Department of Neurology, Boston University School of Medicine, Boston, MA.

   Genomic studies of late-onset Alzheimers disease (LOAD) have identified as many as 22 genes with common susceptibility variants and multiple genes with high-risk, rare coding variants (APP, PSEN2, TREM2, PLD3). While sequencing studies of thousands of individuals remain prohibitively costly, exome arrays, which capture nearly 250,000 putatively functional variants, provide a viable alternative for examining rare coding variants in large samples. Here we present findings from single variant association analyses on 15,788 AD cases and 19,795 controls of European ancestry from the Alzheimers Disease Genetics Consortium (ADGC), Gene and Environmental Risk in Alzheimers Disease (GERAD), and Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE) consortia. The three AD genetics consortia genotyped samples on the Illumina v1.0 (~90%) and v1.1 (~10%) HumanExome arrays, and performed genotype calling and quality control (QC) independently. After QC, data were available for analysis on 30,582 individuals in ten datasets with 203,267 SNPs polymorphic in at least one dataset. We used a score test approach with adjustment for age, sex, and population substructure in each study, and then performed meta-analysis across studies using the R/seqMeta package. Of 16 variants present in more than one dataset and associated with LOAD at P510-8, 12 mapped to the APOE region, including four variants which captured the APOE 2 association and one mapped to the PICALM GWAS signal. The remaining four loci included rare (MAFs=0.002-0.005) missense variants, of which three demonstrated genome-wide significance (from P=2.4710-9 to P=2.2410-8) with large effect sizes (ORs=2.40-4.97). The fourth variant examined was excluded from further consideration, as it demonstrated associations in opposite directions in the two datasets in which genotyping was successful. Additionally, associations of P10-5 were observed at common variants at 11 known LOAD risk loci. Additional genotyping of these and other strongly associated variants (P10-5) is underway in a replication dataset of more than 10,000 samples to confirm these findings. Gene-based analyses are also being performed and will be presented.

You may contact the first author (during and after the meeting) at