of rs12639887 in ADH1B and ADH1C expression or methylation status in brain tissues. However, according to annotation analysis, we can hypothesize that rs12639887 plays a role in determining the differences observed between AAs and EAs in AD GWAS outcomes of ADH1B and ADH1C. In addition to this single case, our data generally revealed that ancestry-related variability in AD GWAS-relevant genes and their surrounding regions can explain some differences in the outcomes of drug dependence GWAS among human populations. Consequently, the differences among ancestry groups in terms of AD genetic predisposition seem not to be influenced by general variation across the genome – that is, these differences may reside in a collection of identifiable ancestry-specific risk alleles. Accordingly, the analysis of local haplotype structure of the PDLIM5-ADH1B-ADH1C region confirmed how population differences in haplotype structure can affect GWAS findings.
