The population structure analysis using all 23 biallelic markers did not find a genetic population substructure (λ = 0.4527) among mixed white Europeans (n = 211), Italians (n = 21) and Portuguese (n = 9). With Structure 2.1, the 3 inferred genetic clusters did not identify the ancestry of these 3 populations. The SNP CHRNB2(rs3926124) was not polymorphic in this population.
The GCF analysis26
revealed a robust correction factor of λ = 0.7245 across the 23 biallelic markers in smokers versus nonsmokers; thus, we did not use a GCF correction for our analyses.
The analysis of smokers versus nonsmokers demonstrates that single genotypes of the 7 markers are not associated with smoking status.
Allelic association was also negative. Sex was strongly associated (p = 0.008), with 71% of men being smokers versus only 47% of women. Age at the time of assessment did not influence smoking status (p = 0.61).
We screened intergenic and intragenic interactions totalling 13 interactions. The interaction between CHRNA4 (rs755203) and CHRNB2(rs1127309) was significantly associated with smoking status (p = 0.003), however, the adjusted p value was not significant (p = 0.72) ().
The 4 markers in α4 were not in linkage disequilibrium (LD), except for CHRNA4(rs755203) and CHRNA4(rs4522666), which showed a strong pairwise LD (p < 0.001, d' = 0.598, r = 0.469).
In CHRNB2, rs1127314 and rs1127309 showed strong LD (p < 0.001, d' = 0.999, r = 0.983).
The global p value for haplotype analysis combining CHRNA4(rs755203) and CHRNA4(rs4522666) was 0.97. The specific haplotype p values were negative as well. The global p value for haplotype analysis of CHRNB2(rs1127314) and CHRNB2(rs1127309) was 0.86.
For the CHRNA4 SNPs that were not in strong LD, we applied the 2-loci interaction analysis as if they were on 2 different chromosomes. The α4 intragenic interaction was significant between CHRNA4(rs755203) and CHRNA4(rs3787116) (p = 0.005), but the adjusted p value was 0.81 ().
When we analyzed the number of cigarettes smoked, we found that the age of the subject was significant (p < 0.001); in particular, age is a risk factor with patients older than 43 years, smoking an average of 25.0 cigarettes daily. Further, the sex of the subject was significant (p = 0.027). Women smoked an average of 16.0 cigarettes daily, whereas men smoked 21.0 cigarettes daily. The SNP CHRNA4(rs3746372) was found to be significant (p = 0.01), and those with genotype (1,1) smoked 22.0 cigarettes daily, whereas the other patients smoked an average of 17.0 cigarettes daily.
In regard to the allele effect, we found that allele 1 (frequency 74.3%) had a specific effect (p = 0.02), with a mean number of 20.42 cigarettes smoked daily, compared with allele 2, which was associated with a mean number of 17.07 cigarettes smoked daily.
CHRNA4(rs3787116) showed a slight trend (p = 0.059), with the heterozygous genotype (1,2) protecting against heavy smoking (average 18.0 cigarettes daily), whereas subjects with homozygous genotypes smoked 21.0 cigarettes daily.
When we employed the Helix Tree 2-loci interaction analysis, we did not include the uninformative SNP CHRNB2 (rs3926124). The interaction of SNP CHRNA4(rs755203) and SNP CHRNB2(rs1127309) was significant (p = 0.07) (). However, after the multiple test correction, the adjusted p value became insignificant (p = 0.06). Surprisingly, both SNPs were not significant on their own.
The global p value for the haplotype analysis of CHRNA4 (rs755203) and CHRNA4(rs4522666) was 0.98. The global p value for the haplotype analysis of CHRNB2 (rs1127314) and CHRNB2(rs1127309) was 0.77.
CHRNA4(rs3787116) and CHRNA4(rs3746372) showed a significant interaction (p = 0.004). Further, the adjusted p value was also significant (p = 0.03). The interaction between CHRNA4(rs3787116) and CHRNA4(rs4522666) was significant (p = 0.001), as was the adjusted p value (0.001). Like-wise, the interaction between CHRNA4(rs3746372) and CHRNA4(rs4522666) was significant (p = 0.004), as was the adjusted p value (0.033).