We focused on markers that either overlapped between the GIANT consortium and this study or GIANT consortium markers that were in reasonable linkage disequilibrium (r2> 0.80)with SNP markers from this study. In all, we identified 717,411 markers to be tested for association that also have corresponding weights from the GIANT consortium. As a result, p-values from this study may be up-weighted or down-weighted depending upon the association signal from the GIANT GWAS. More specifically, the unweighted (nominal) p-values from this study are divided by the weights as assigned through the GIANT GWAS to generate the weighted p-values. The weighting procedure was conducted in R using the “weighted_FDR.R” script that can be found at http://www.wpic.pitt.edu/wpiccompgen/fdr/.
