Poster Presentation 9th GeneMappers Conference 2012

Using meta-analysis data from GWAS to predict truly replicable risk alleles for Parkinson’s disease: Experience from the GEO-PD consortium. (#112)

George D. Mellick 1
  1. Griffith University, Nathan, QLD, Australia

Recent GWAS studies of Parkinson’s disease (PD) have identified loci which modestly influence risk for the disease and may prove valuable for predicting patient risk and improving the etiological understanding. Meta-analytical methods have now been used in an attempt to identify addition loci through imputation of original data or via direct analysis of publically available datasets [1]. These efforts have revealed many potential genetic risk factors that now require replication in large population samples worldwide to establish their true utility. The Genetic Epidemiology of Parkinson’s Disease (GEO-PD) consortium is a replication engine ideal for the purposes of replication studies. GEO-PD now includes 50 sites from 25 countries, sharing DNA from 34,000 PD cases and 30,859 controls. Recently the GEO-PD examined twenty loci (each represented by one SNP), assessed by meta-analysis as those with “strong epidemiologic credibility” for association [1]. DNA from 17705 individuals was studied (8750 cases and 8955 controls from 21 sites including 2024 Australians). Genotyping using the Sequenom platform was performed at core facilities (Department of Human Genetics, Helmholtz Zentrum, Munich). Fixed as well as random effects models were used to provide the summary PD risk estimates for these variants. The between-study heterogeneity and the influence of population ancestry was also examined. Nine of the examined SNPs were strongly associated with PD in the replication analysis confirming the findings for SNCA, LRRK2, MAPT, BST1, GAK, STK39, LAMP3, HIP1R, SYT11/RAB25. Evidence for association was also obtained for PARK16 but several others (including HLA-DRB5 and ACMSD) did not replicate.  These data provide an interesting opportunity to examine the properties of the initial meta-analysis results and the replication sample that may assist in predicting the replication potential of GWAS data and the selection of candidates to prioritize in subsequent examinations.

[1] Lill et al. PLoS Genetics 2012, 8(3), e1002548