In this study, we have selected subjects as smokers if they had ever smoked, as reported on the smoking initiation history. We used both former and current smokers in the GWAS and the replication study according to smoking cessation history, with a proper treatment in the linear regression model (see statistical methods below). We excluded subjects with invalid CPD information ( e.g., 5–10 cigarettes/day or 5 cigarettes/day). We also exclude subjects below the age of twenty or who did not provide their sex.