In this study, we used a quantitative design for both VAANX-ND and VAFTND samples. The distribution of FTND scores in the two samples are plotted in Supplementary Figure S1 in the Supplementary Material. The mean FTND score was slightly lower in the VAFTND sample (mean, s.d., 4.27, 2.62) than in the VAANX-ND sample (5.18, 2.57). In the VAANX-ND sample, rs16969968 and rs1051730 were nominally significant (Table III), but only rs16969968 survived correction for multiple testing given the effective number of independent tests estimated at 5 by the SNPSpD method. In VAFTND, four of the seven markers showed nominal significance (Table III), rs16969968 and rs1051730 remained significant after Bonferroni correction. In both samples, the same alleles were over-represented in those subjects with higher FTND scores. For rs16969968, the minor allele A (Asn) was more frequent in those with higher FTND scores; for rs1051730, it was also the minor allele (the A allele) that was over-represented in subjects with higher FTND scores. For both markers, the risk alleles were the same as reported in previous studies [Saccone et al., [2007]; Berrettini et al., [2008]; Thorgeirsson et al., [2008]].
