There are several factors contributing to false positives in case control studies, including population stratification, sample size and genotyping error. In our study, we have taken measurements to reduce these risks. We genotyped unlinked microsatellite markers to assess potential stratification and used reasonable sample sizes. We examined HWE for all SNPs. For those two SNPs showing departure from HWE, we checked each subject groups separately. Only one (rs10485057) of these two SNPs involves in the core risk haplotype. For this marker, HWE deviation is caused largely, if not exclusively, by the High-ND smokers. This case-specific HWE deviation, by itself, can be interpreted as an association [34,35]. Multiple testing is another source of false positives. In this study, none of the tests would survive Bonferroni correction.
