narrowly missed significance when the 358 SNPs examined within the gene were simultaneously considered (pcorrected=0.060). None of the SNPs that were associated with externalizing at the p<0.01 level were associated with PTSD (all p>0.15). When we examined ANK3 for association with PTSD, the most significant result was observed at three SNPs in perfect LD, all with p=0.00060 (Table 1, SNPs 6 to 8). These SNPs remained significant after multiple-testing correction (pcorrected= 0.045). Other SNPs with p<0.01 were observed over a 69 kb interval at chr10:61,799,888-61,869,059 which includes the location of rs9804190. The bipolar disorder-associated SNP rs9804190 was the only SNP in ANK3 that was nominally significantly associated (p<0.05) with both PTSD and externalizing. No SNPs with p<0.01 were observed outside of the boundaries of the brain-specific isoforms for either trait. This included the two genotyped bipolar disorder risk SNPs in the 5′ region, rs10994336 and rs10994397 (externalizing p=0.31 and p=0.28 respectively, PTSD p=0.48 and p=0.27 respectively).
