Unadjusted MAFs and the results of association analyses for OPRD1 SNPs comparing cases to non-dependent controls are displayed in Table 1. Associations of at least nominal significance are seen for 10 of the 21 OPRD1 SNPs. For 4 SNPs (for which pairwise r2 ranged from 0.68–1.0, see Figure 1), the p values are less than the significance threshold calculated to correct for multiple comparisons. Based on the LD relationships of the OPRD1 SNPs, it appears likely that these results could largely be explained by a single strong association signal (minimal p value = 2.95 × 10−4 for rs2236857 and rs2236855). To test this assertion, association for those OPRD1 SNPs of at least nominal significance was reexamined, conditioned on rs2236857 allelic dosage. The only residual signal observed was for rs581111, for which only mild attenuation was found (p=0.059), consistent with the lack of LD (r2=0) for this SNP with either rs2236857 or rs2236855. A post-hoc two locus association analysis of rs2236857 and rs581111 (Table 2) found that the GA haplotype consisting of the coupled minor alleles (prevalence 0.077) is more strongly associated with an odds ratio (OR) of 1.68 (p=1.41 × 10−5).
