A Tutorial of Meta-Analysis for Genome-wide Association Studies

Statistical methods for meta-analysis of genome-wide association studies will be reviewed. Emphasis will be placed on the random model approach to meta-analysis, in which the estimated marker effect of each locus from each study will be treated as a random effect following a normal distribution with an unknown mean (main effect) and an unknown variance. Likelihood ratio and score tests will be used in the study. Rejection of the null hypothesis of zero mean and zero variance indicates a positive detection of the association of the current marker with the trait of interest.