rs12471739 at NR4A2 had significant cis-acting regulatory effects on the transcript expression of NR4A2 in the PBMC (p<0.048) (Table 5). For KCNB2, one SNP rs2929576 with p = 2.71×10−7 in the meta-analysis was confirmed in the family sample (p = 0.0142). Furthermore, five flanking SNPs were located in KCNB2 on chromosome 8 (Table 4). rs2958414 at KCNB2 had significant cis-acting regulatory effects on the transcript expression of KCNB2 in the brain (p<0.018) (Table 5). In addition, several flanking SNPs demonstrated borderline associations with MaxDrinks that have been listed in the supplement Table S1. For example, rs7624305 (p = 1.06×10−1) near SGOL1, rs2620418 (p = 1.03×10−2) and rs11938588 (p = 2.03×10−2) in NDST4, rs2620418 (p = 2.09×10−2) in DTWD2, rs7624305 (p = 1.06×10−1) in 3p24.3 and rs4947510 (p = 9.16×10−2) in DDC. However, the top hit SNP from meta-analysis (rs11128951) was not confirmed in the family sample possibly due to cause of the ambidirectional effect indicated by the significant Cochrane’s Q value (p=0.04). The cis-eQTL signal of this top hit SNP was not detected either. Besides, all other signals have the same direction of effect (p ≥ 0.05).
