In the GWAS and the replication study, the association of each CNP or SNP with CPD was assessed using a linear regression model that assumed additive effects for copy number dosage or allele dosage (ranging from 0.0 to 4.0 for CNPs and from 0.0 to 2.0 for SNPs, respectively). In the regression model, gender, age, age-squared, smoking cessation status (former/current), status for the 19 diseases and the first and second principal components (PCs) in the East Asian population (Figure S2; see reference [26] for details of the Japanese population structure) were adopted as covariates. In the replication study, the most significantly associated CNPs and SNPs in the loci that had in the GWAS were evaluated using R statistical software. For the CNP and the SNPs in the combined GWAS and replication study, associations were evaluated using the results of the genome-wide meta-analysis for CPD. The combined results of the studies were obtained using an inverse-variance method from the summary statistics (effect size) and the standard error.
