Table 1 shows the number of supplementary SNPs needed to tag all common variants in various populations for our primary set of 910 genes that are biologically relevant to addiction. This set includes 486 genes that were derived mainly through an expert nomination process, and 424 additional genes that correspond to roughly the top 5% of genes identified using mouse systems genetics (Chesler and colleagues, submitted). Together this set of 910 genes was our primary set for the analysis of microarray coverage (see supporting file S1 for the complete list of these genes). We assessed the SNP coverage of these genes by determining if common SNPs (MAF≥5%) were tagged through LD by SNPs on a particular microarray. In Table 1, for each array and each population, we report the number of common SNPs in these genes that fail to satisfy r 2≥0.8 with a SNP from the array; that is, the number of SNPs not tagged by the array. For example, we found that 57% of the common SNPs in these genes were not tagged by the Affymetrix 5.0 SNP
