of a single haplotype block, it has been challenging to determine which SNP(s) in which gene(s) cause the association and whether the association is due to coding differences (e.g., rs16969968 which causes D398N substitution in CHRNA5) or expression differences, which might be due to either coding or non-coding SNPs. Indeed, statistical analyses suggest that there is likely to be more than one functionally significant allele at this locus (Saccone et al. 2010). Animal models have been used to address this issue. Fowler et al. (2011) observed a large increase in nicotine self-administration in mice that were homozygous for null alleles of the Chrna5 gene. Importantly, they were able to reverse this effect by replacing Chrna5 expression in the habenula, and to replicate the effect of the null allele by knocking-down Chran5 in the habenula, suggesting that Chrna5 expression in this region was necessary and sufficient for normal nicotine self-administration behavior. Similarly, Frahm et al. (2011) showed that overexpression of Chrnb4 decreased nicotine aversion and that this effect was reversed by viral-mediated expression of the Chrna5 D398N variant in the habenula. Thus, mouse models have helped to elucidate possible roles for two of the genes in the haplotype. Both of these
