Genotype imputation methods [21],[14] are now being widely used in the analysis of genome-wide association studies [5] and meta-analysis of such studies [22],[20]. These methods can be thought of as a more sophisticated version of Multi Marker tests but are relatively much more computationally demanding. We carried out an evaluation of the boost in power that can be gained by imputation using the program IMPUTE [14]. For our simulations with a sample size of 2000 cases and 2000 controls and a relative of the causal SNP or 1.3 we ran IMPUTE on the genotype data from each of the chips under study using the CEU HapMap as the basis for imputation. We then carried out a test of association at all the imputed SNPs in addition to the SNPs on each chip. We used our program SNPTEST to carry out tests of association at imputed SNPs to properly account for the uncertainty that can occur at such SNPs[14]. The results of the simulations are shown in Table 1 and shows that the use of IMPUTE provides a noticeable boost in
