genome-wide association analyses, we set genome-wide significance as p<5 × 10−8 and suggestive significance as 5 × 10−8<p<5 × 10−7. For other analyses, we used a Bonferroni correction for multiple testing. The appendix (p 17) describes quality control after association testing. For the lead single nucleotide polymorphism (SNP) at each of our novel signals of association with FEV1 extremes, we tested for association with COPD risk using the aforementioned definition (appendix p 18). We did a meta-analysis across smoking strata using inverse variance weighting. We assessed evidence for polygenic architecture of FEV1-defined phenotypes (appendix p 18).33 For this analysis, we created discovery and target subpopulations, each of which comprised cases and control groups created by randomly splitting the low FEV1 and average FEV1 groups (appendix pp 18–20). Variants of MAF of at least 1% associated with low FEV1 below given p value thresholds in the discovery population were incorporated into an aggregate score, and the association with the aggregate score was tested in the independent target population. A similar approach (appendix pp 18–20), in each case using independent discovery and target populations, was used to test for a shared polygenic component between high FEV1 and low FEV1, low FEV1 in
