list of hits from the GWAS associations can still be genuine associations. For definiteness, we focus throughout on the threshold of p<5×10−7). This is deliberately set so that false positive rates will be low – for example, most SNPs with trend test p-values passing this threshold in GWAS studies, including all of those in the WTCCC experiment, have had associations confirmed in replication studies (see [18] and the NHGRI Catalog of Published Genome-Wide Association Studies at http://www.genome.gov/GWAStudies/). Choice of a different p-value threshold changes the numerical value of the power we calculate, but does not affect the relative performance of the chips, or the relative effect of sample size (data not shown).
