None of the top SNPs (or associated genes) identified in our study (at P < 5×10−5) overlapped with any of the top 54 markers reported by Anney and colleagues to be associated with CD traits in ADHD probands at P < 10−5 Furthermore, combined examination of the list of genes from our sample and the Anney sample, using the databases indicated above, did not indicate that the findings were enriched for any gene ontology category. However, three genes from the combined list (PDX1, ATP8B1 and ERCC4) have been implicated in pancreas abnormalities; these results are significant (FDR= 0.000047) and might be relevant given the role of the pancreas in the endocrine system. However, the largely nonoverlapping findings from the studies likely reflects differences in the sample ascertainment. Anney and colleagues studied CD traits among individuals with ADHD. Furthermore, the average age of that sample was 10.88 years (s.d. = 2.8). Our sample was ascertained through substance-dependent probands, although phenotypic information was available on the controls, in addition to the cases, which enabled us to study CD symptomatology among individuals with
