Genome-wide meta-analyses identify multiple loci associated with smoking behavior

Abstract

Consistent but indirect evidence has implicated genetic factors in smoking behavior. We report meta-analyses of several smoking phenotypes within cohorts of the Tobacco and Genetics Consortium (n = 74,053). We also partnered with the European Network of Genetic and Genomic Epidemiology (ENGAGE) and Oxford-GlaxoSmithKline (Ox-GSK) consortia to follow up the 15 most significant regions (n > 140,000). We identified three loci associated with number of cigarettes smoked per day. The strongest association was a synonymous 15q25 SNP in the nicotinic receptor gene CHRNA3 (rs1051730[A], beta = 1.03, standard error (s.e.) = 0.053, P = 2.8 x 10(-73)). Two 10q25 SNPs (rs1329650[G], beta = 0.367, s.e. = 0.059, P = 5.7 x 10(-10); and rs1028936[A], beta = 0.446, s.e. = 0.074, P = 1.3 x 10(-9)) and one 9q13 SNP in EGLN2 (rs3733829[G], beta = 0.333, s.e. = 0.058, P = 1.0 x 10(-8)) also exceeded genome-wide significance for cigarettes per day. For smoking initiation, eight SNPs exceeded genome-wide significance, with the strongest association at a nonsynonymous SNP in BDNF on chromosome 11 (rs6265[C], odds ratio (OR) = 1.06, 95% confidence interval (Cl) 1.04-1.08, P = 1.8 x 10(-8)). One SNP located near DBH on chromosome 9 (rs3025343[G], OR = 1.12, 95% Cl 1.08-1.18, P = 3.6 x 10(-8)) was significantly associated with smoking cessation.

Figure 1

Genome-wide association results for the TAG Consortium. Manhattan plots showing significance of association of all SNPs in the TAG Consortium meta-analyses for four smoking phenotypes. (a–d) Manhattan plots show SNPs plotted on the x axis according to their position on each chromosome against, on the y axis (shown as −log₁₀ P value), the association with CPD (a), former versus current smoking (b), ever versus never smoking (c) and age of smoking initiation (d).

Figure 2

Forest and regional plots of significant associations for CPD from meta-analyses of the TAG, Ox-GSK and ENGAGE consortia. (a–f) Regional association plots show SNPs plotted by position on chromosome against −log₁₀ P value with each smoking phenotype. Estimated recombination rates (from HapMap-CEU) are plotted in light blue to reflect the local LD structure on a secondary y axis. The SNPs surrounding the most significant SNP (red diamond) are color coded to reflect their LD with this SNP (using pairwise r ² values from HapMap-CEU): blue, r ² ≥ 0.8–1.0; green, 0.5–0.8, orange, 0.2–0.5; gray, <0.2. The gray bars at the bottom of the plot represent the relative size and location of genes in the region.

Figure 3

Forest and regional plots of significant associations for smoking behavior. (a–d) Shown are plots for smoking initiation (a,b) and smoking cessation (c,d) from meta-analyses of the TAG, Ox-GSK and ENGAGE consortia. Regional association plots show SNPs plotted by position on the chromosome against −log₁₀ P value with each smoking phenotype. Estimated recombination rates (from HapMap-CEU) are plotted in light blue to reflect the local LD structure on a secondary y axis. The SNPs surrounding the most significant SNP (red diamond) are color coded to reflect their LD with this SNP (using pairwise r² values from HapMap CEU): blue, r² ≥ 0.8–1.0; green, 0.5–0.8; orange, 0.2–0.5; gray, <0.2. The gray bars at the bottom of the plot represent the relative size and location of genes in the region.