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 heavy smokers and in never smokers, and low FEV1 in participants who did and those who did not report a history of doctor-diagnosed asthma. To show the reliability of the doctor diagnosis of asthma variable, we showed association with asthma at ten previously reported genome-wide significant loci (appendix pp 21, 29).
