We sought to link our novel results in mice to human alcoholism using data from COGA. We selected two phenotypes, DSM-IV AD and alcohol withdrawal, that were related to preference measured in the mouse model. AD is characterized by excessive intake on a regular basis, while withdrawal reflects negative consequences from drastic reductions in alcohol intake. A total of 43 SNPs in PPARA were genotyped; four provided evidence of association with withdrawal (5.1 × 10−3<p<0.04; Figure 3A) while none were associated with AD (p>0.15; data not shown). A total of 107 SNPs were tested in PPARG; five SNPs provided evidence of association with withdrawal (9.5 × 10−3<p<0.05; Figure 3B) and 1 with AD (p=0.03; data not shown). None of the 30 SNPs in PPARD supported an association with either AD (p>0.22) or withdrawal (p>0.38) (Supplemental Figure 5). We extended our studies of PPARG to include the gene for its transcriptional coactivator, PPARGC1A. We tested 46 SNPs and 3 provided support for an association with AD (8.6 × 10−3<p< 0.04) but none with withdrawal (p>0.07) (Figure 3C). In all regions with
