There are several limitations of our study. Because this is a cross-sectional study and these analyses rely on self report and are subject to recall bias, we assumed no temporal sequence in our studies of comorbidity. Our sample size may be sufficient to test main effect associations, but power is quite limited to detect any gene-comorbidity interactions of moderate effect size, so the results must be interpreted with caution. An estimated sample of 10,000-20,000 is required for such interaction analyses. Lastly, our sampling selection for cases and controls was based on FTND scores, and we need to be cautious about the interpretation of the relationship of comorbid psychiatric disorders and nicotine dependence over the full range of smoking behaviors. In addition, comorbid psychiatric illness defined in this study does not represent all psychiatric disorder found in the community, but represents a subset of individuals who smoke cigarettes and have lifetime psychiatric illness. Thus, further examination of the relationship between psychiatric disorders and nicotine dependence needs to be done in different community based samples to increase generalizability.
