Additive (AA versus AB versus BB) model was used for the tests of association by genotype and diplotype. Diplotype is defined as a specific combination of two haplotypes. The statistical analyses were performed using PLINK version 1.07 (http://pngu.mgh.harvard.edu/~purcell/plink).

Haploview 4.2 (http://www.broad.mit.edu/mpg/haploview/) was used with Gabriel’s rule to determine the haplotype and linkage equilibrium (LD) structure of the ALOX5AP gene. The SNP rs9506352 associated significantly with FEV1 when the Ansung data were examined separately or combined data Sorafenib concentration [P = 0.009 and 0.006 (permuted P = 0.045 and 0.032), respectively]; FEV1 increased by 2.616 and 1.246 per the minor A allele was present, respectively. The SNP rs10162089 and rs3803277 were significantly associated with FEV1 in combined data (P = 0.027

and 0.011), FEV1 increased by 0.968 and 1.008 per the minor A and C allele was present, respectively. In contrast, FEV1/FVC did not associate significantly with any of the SNPs in the Ansan, Ansung, or total populations. Table 2 indicates the associations between the SNPs in the ALOX5AP and FEV1 or FEV1/FVC. Two LD blocks were identified among the 13 intronic SNPs in the ALOX5AP gene (Fig. 1). The haplotypes with frequencies below 5% were filtered out. Ten SNPs were included in the second LD block, which had a relatively high D’ (>0.9) and R2 value as well as containing two exons. Therefore, diplotypes with tagging check details SNPs were used for analysis. Each LD block had three and four haplotypes, respectively. Of these, the diplotype of haplotype AA in block 1 associated significantly with FEV1 (P = 0.023); FEV1 increased 0.997 per haplotype AA was existed. The diplotype

of haplotype TCAC in block 2 also associated significantly with FEV1 (P = 0.008 and permuted P = 0.044); FEV1 increased by 1.230 per haplotype TCAC was present. FEV1/FVC did not associate with any diplotypes. Table 3 indicates the associations between the diplotypes in the ALOX5AP and FEV1 or FEV1/FVC. The SNP rs9579648 was associated with FEV1 in Ansan data (P = 0.044); mafosfamide FEV1 decreased by 2.660 per the minor G allele was present. Except rs9579648, SNPs in ALOX5AP showed no significant interaction with smoking on both FEV1 and FEV1/FVC. (Data not shown). In the results of analysis for general population (8535 subjects), for one minor allele of rs10162089, FEV1 was 1.135 and 0.622 higher as compared to wild type carriers in Ansung and combined data (P = 0.023 and 0.041, respectively). The SNPs rs9506352 was associated with decreased FEV1 in Ansung and combined data (P = 0.020 and 0.019, respectively); FEV1 increased by 1.225 and 0.749 per the minor A allele was present. For one minor allele of rs3803277, FEV1 was 1.224 and 0.823 higher as compared to wild type carriers in Ansung and combined data (P = 0.007 and 0.003 (permuted P = 0.033 and 0.014), respectively).