Polygenic score (PGS) analyses of individual criteria in NESARC-III revealed an interesting pattern of associations. First, none of the FTND-specific items were correlated with the DSM-NicDep PGS, suggesting that the DSM-NicDep PGS captured genetic variance that was distinct from the FTND items. Second, even though DSM-IV did not include a diagnosis of nicotine abuse, DSM-5 includes 3 abuse criteria in its diagnosis of Tobacco Use Disorder – of these, two were unrelated to the DSM-NicDep PGS (Social, Fail) while the third (Hazard) was significantly associated. Notably, Fail was unrelated to both PGS and was amongst the least commonly endorsed criteria. Third, while the FTND is presumed to evaluate aspects of both tolerance and withdrawal, the DSM Tolerance criterion was unrelated to the FTND PGS. In contrast, Withdrawal as assessed by the DSM was associated with both PGS, while Time to first cigarette was solely associated with FTND. Taken together, these item-level analyses suggest that the DSM-NicDep GWAS may have indexed genetic liability to a distinct set of tobacco-related problems than the FTND GWAS. However, many confidence intervals on these estimates
