differing in both size and location. Polymorphic CNVs, which are deletions/amplifications that are common in a population (above 1% frequency), can be used in tests of association, just like SNPs. We hypothesized that CNVs could be significantly associated with smoking-related phenotypes. However, because CNVs are based on estimated deletions/amplifications, inferences based on significant association tests can be more tenuous, although they can still identify genes potentially influencing complex phenotypes, such as quantitative measures of smoking behavior.
