Our understanding of the molecular genetic contributions to smoking are limited to the additive effects of individual single nucleotide polymorphisms (SNPs), but the underlying genetic risk is likely to also include dominance, epistatic, and gene-environment interactions. To begin to address this complexity, this study attempted to identify potential genetic interactions between rs16969968, the most replicated SNP associated with smoking quantity, and all SNPs and genes across the genome. Using the UK Biobank, we found one gene,PCNA, that showed a genome-wide significant interaction with rs16969968 for smoking behaviors in a sample of 116 442 smokers of European ancestry. We replicated this finding in a meta-analysis of five Finnish samples (n = 40 140): FinHealth, FINRISK, Finnish Twin Cohort, GeneRISK, and Health-2000-2011. To our knowledge, this represents the first reliable epistatic effect between measured genetic variants for smoking behaviors and provides a novel direction for possible future functional studies related to this interaction. Furthermore, this work demonstrates the feasibility of these analyses, which may be applied to other top SNPs for smoking and/or other phenotypes.