We imputed the genotype data 1Mb surrounding the most significant gene region that covers the entire NKAIN1-SERINC2 region, to fill in the missing markers (see imputation below). Associations between all available markers within this 1Mb region and alcohol dependence were tested in Australians using a family-based association test program (FBAT; Horvath et al., 2001), adjusting for covariates including sex, age and admixture rates (ancestry proportions; see above), and assuming an additive genetic model under the null hypothesis of no linkage and no association, biallelic mode, minimum number of informative families of 10 for each analysis and offset of zero (results same as those using the program DFAM implemented in PLINK; data not shown). The associations that were replicated across European-American discovery samples and Australians are shown in Table 2 and Figure 1. This replication design helped to control for the false positive findings. Meta-analysis was performed to derive the combined p values between European-American discovery sample and Australians using the program METAL. An overall z-statistic and p-value for each SNP were calculated from a weighted sum of the individual statistics.
