The discovery PheWAS was conducted in the WHI nicotine-exposed sample using Plink 1.922. Logistic and linear regression analyses were used to calculate the association between genetic variants and phenotypes (binary and quantitative, respectively). The regression models included as covariates age, age-squared, and the first 10 principal components to control for differences in ancestry. Since PheWAS are not discovery studies (they are follow-up investigations useful to delineate the role of previously identified loci in the human phenome), they can be corrected with less stringent multiple testing criteria than the Bonferroni correction for the number of independent tests3, 26, 27. We adjusted our p values using a locus-wise Bonferroni correction (which corresponds to correction for the number of phenotypes studied) that was recently proposed by Simonti and colleagues27. Accordingly, we calculated that the phenome-wide significance threshold to keep the type I error rate at 5% is p = 1.40 × 10−4. Since the LLFS cohort is ~40% smaller than WHI cohort (i.e., the association analysis in the replication cohort had Lower statistical power than that conducted on the discovery cohort), we considered
