S1‐03‐05: U.S. Alzheimer's Disease Genetics Consortium: Organization and results

not available. S1-03-03 THE X-CHROMOSOME AND ALZHEIMER’S DISEASE Steven Younkin, Mayo Clinic College of Medicine, Jacksonville, FL, USA. Contact e-mail: younkin.steven@mayo.edu Abstract not available.not available. S1-03-04 FAMILY-BASED APPROACHES TO ANALYSIS OF GWAS DATA Ellen M. Wijsman, University of Washington, Seattle, WA, USA. Contact e-mail: wijsman@u.washington.edu Abstract not available.not available. S1-03-05 U.S. ALZHEIMER’S DISEASE GENETICS CONSORTIUM: ORGANIZATION AND RESULTS Lindsay A. Farrer, Boston University School of Medicine, Boston, MA, USA. Contact e-mail: farrer@bu.edu Background: The Alzheimer Disease Genetics Consortium (ADGC) was formed to use collaboratively the collective resources of the AD research community in the United States to identify AD genes using the genome wide association study (GWAS) approach. Methods: The ADGC includes nearly all of nation’s leading scientists who are studying the genetic basis of AD in human populations, as well as experts in the clinical, neuropathological, molecular and statistical aspects of AD. The ADGC capitalizes on the vital and varied resources of the 29 National Institute on Aging AD Centers, several large longitudinal and cross-sectional studies of cognitive decline and dementia, and the National Cell Repository for AD. The primary goal of the ADGC is to identify genetic risk factors for AD and onset age, a clinical milestone that is related to rate of progression through the prodromal and mild cognitive impairment (MCI) phase of the disease. A secondary goal is to identify genetic variation that influences specific AD-related endophenotypes such as neuropathology features (e.g. plaque load, tangle distribution, etc), biomarker measures [e.g. cerebral spinal fluid (CSF) Ab and tau levels], rate-of-cognitive decline (i.e., the clinical hallmark of AD), and responses to environmental factors (e.g., drugs, education). The ADGC promotes discovery by encouraging analyses using novel statistical, bioinformatic and system biology approaches. Results: Thus far, 17 GWAS datasets comprising more than 11,000 AD cases and 13,500 cognitively normal controls aged 60 years and older have been assembled for consortium analyses. Nearly 1,100 of these subjects are African American and 1,100 are Hispanic. Analysis of data from 5,615 AD cases and 5,778 controls (all Caucasian) has confirmed associations with CR1 (e.g. rs3818361, OR 1⁄4 1.15, CL 1⁄4 1.071.23, p 1⁄4 7.5x10-5), CLU (e.g. rs11136000, OR 1⁄4 0.91, CL 1⁄4 0.85-0.96, p 1⁄4 0.0014), and PICALM (e.g. rs561655, OR 1⁄4 1.15, CL 1⁄4 1.09-1.23, p1⁄4 5.4x10-6) that were reported previously by the UK and French AD consortia. There was significant evidence for interaction of PICALM and APOE on AD risk due primarily to a much stronger association with PICALM among subjects having the e4 allele. Conclusions: We successfully replicated associations with CR1, CLU, and PCALM. Infrastructure is in place for discovery of the genetic basis of AD and related endophenotypes. S1-03-06 IS IT REAL? THE SHORT HISTORY OF GENOMEWIDE ASSOCIATION FINDINGS IN ALZHEIMER’S DISEASE Lars Bertram, Max-Planck Institute for Molecular Genetics, Berlin, Germany. Contact e-mail: bertram@molgen.mpg.de Background: Within the past two years, over one dozen independent genome-wide association studies (GWAS) have been published in the field of Alzheimer’s disease (AD), highlighting over thirty novel potential susceptibility loci beyond the well-established APOE association. Currently, the most compelling GWAS signals are observed in CLU (APOJ), PICALM, CR1, and GAB2. The vast majority of these signals emerged from heterogeneous multicenter case-control studies, which lack independent replication in samples ascertained from AD families. Methods: The purpose of this study was to assess the most compelling AD GWAS signals in a collection of nearly 1,250 independent AD families. Overall, we directly genotyped over 40 single-nucleotide polymorphisms (SNPs) across 25 different GWAS loci in these samples. In addition, we tested these SNPs for their effect on AD risk and CSF Abeta and tau levels in an independent case-control dataset (n 1⁄4 455), not previously studied for any of these markers. All data were then combined with the publicly available genotypes by random-effects meta-analysis. Results: Association analyses confirmed the previous associations between risk for AD and variants in CLU, PICALM, CR1, and GAB2 (P-values ranging from 0.0005 to 0.05). Notably, these associations were found with the same alleles and the same direction of effect as in the original reports, in some cases substantially strengthening the overall evidence by meta-analysis (P-values now ranging from 1x10-7 to 1x10-16). In addition, several other GWAS loci were found to be associated in our study, albeit less consistently and with weaker statistical support. Consistent with their effects on disease risk, some alleles also showed association with CSF Abeta and tau levels in our case-control dataset. Conclusions: The independent convergence of case-control GWAS and family-based follow-up findings substantially strengthens the notion that variants in CLU, PICALM, CR1, GAB2 and a few other genes exert genuine effects on AD pathogenesis. Furthermore, the results from our CSF biomarker analyses provide preliminary insights regarding the predominant pathogenetic mechanisms underlying some of these genetic associations.