Our approach to analyzing susceptibility pathways for smoking quantity was based on the identification of GO terms and KEGG pathways common to the OZALC-NAG and SAGE datasets that showed replication in the ARIC study. Our decision to unify the set of SNPs analyzed in each of the two exploratory studies, genotyped by different Illumina chips, proved to be a valid option that identifies common categories of genes while maximizing the chances of observing the same enriching genes. We implemented this strategy by extending the set of SNPs originally genotyped in the OZALC-NAG study, incorporating imputed data to encompass the ones ascertained in SAGE. We applied this same approach to analyze the ARIC dataset, which was genotyped using the Affymetrix platform. Moreover, apart from the differences in the enriching genes, all of the replicated GO terms and KEGG pathways were also significant when we restricted analysis of ARIC to the genotyped SNPs in the Affymetrix Human SNP Array 6.0 (data not shown).
