For Manuscript Submission, Check or Review Login please go to Submission Websites List.
For the academic login, please select your country in the dropdown list. You will be redirected to verify your credentials.
On the Meta-Analysis of Genome-Wide Association Studies: A Robust and Efficient Approach to Combine Population and Family-Based StudiesWon S.a, b · Lu Q.c · Bertram L.d, e · Tanzi R.E.d · Lange C.f–j
aDepartment of Applied Statistics, and bThe Research Center for Data Science, Chung-Ang University, Seoul, Republic of Korea; cDepartment of Epidemiology, Michigan State University, East Lansing, Mich., dGenetics and Aging Research Unit, Department of Neurology, Massachusetts General Hospital, Charlestown, Mass., USA; eNeuropsychiatric Genetics Group, Department of Vertebrate Genomics, Max Planck Institute for Molecular Genetics, Berlin, Germany; fDepartment of Biostatistics, Harvard School of Public Health, gHarvard Medical School, and hCenter for Genomic Medicine, Brigham and Women’s Hospital, Boston, Mass., USA; iInstitute for Genomic Mathematics, University of Bonn, and jGerman Center for Neurodegenerative Diseases (DZNE), Bonn, Germany Corresponding Author
Department of Applied Statistics
Seoul 156-756 (Republic of Korea)
Tel. +82 2 820 5217, E-Mail email@example.com
For the meta-analysis of genome-wide association studies, we propose a new method to adjust for the population stratification and a linear mixed approach that combines family-based and unrelated samples. The proposed approach achieves similar power levels as a standard meta-analysis which combines the different test statistics or p values across studies. However, by virtue of its design, the proposed approach is robust against population admixture and stratification, and no adjustments for population admixture and stratification, even in unrelated samples, are required. Using simulation studies, we examine the power of the proposed method and compare it to standard approaches in the meta-analysis of genome-wide association studies. The practical features of the approach are illustrated with a meta-analysis of three genome-wide association studies for Alzheimer’s disease. We identify three single nucleotide polymorphisms showing significant genome-wide association with affection status. Two single nucleotide polymorphisms are novel and will be verified in other populations in our follow-up study.
© 2012 S. Karger AG, Basel