PMC

a, Stage 1; b, stage 2. Black dots are the uncorrected test statistics. Red dots are the statistics corrected by the genomic control method (λ=1.03 for stage 1, λ=1.06 for stage 2). Under the null hypothesis of no association at any locus, the points would be expected to follow the black line.

a, rs2981582; b, rs3803662; c, rs889312; d, rs13281615; and e, rs3817198. The x-axis gives the per-allele odds ratio. Each row represents one study (see ), with summary odds ratios for all European and all Asian studies, and all studies combined. The area of the square for each study is proportional to the inverse of the variance of the estimate. Horizontal lines represent 95% confidence intervals. Diamonds represent the summary odds ratios, with 95% confidence intervals, based on the stage 3 studies only.

a, Map of the whole FGFR2 gene, viewed relative to common SNPs on HapMap. The gene is 126 kb long and in reverse 3′–5′ orientation on chromosome 10. Exon positions are illustrated with respect to the 67 SNPs with m.a.f.>5% in HapMap CEU (therefore the map is not to physical scale). Numbered SNPs are those tested in the genomewide study. SNPs in black were not significant in stage 1. Those in red were significant at P<0.0001 after stage 2. rs10510097 (orange) was significant in stage 1, but failed quality control in stage 2 owing to deviation from Hardy–Weinberg equilibrium. Squares indicate pairwise r² on a greyscale (black=1, white=0). Red circle indicates rs2981582. b, Resequenced 32 kb region, shown relative to SNPs in CEU with m.a.f.>5%, showing pairwise LD for SNPs in HapMap CEU (left panel) and JPT/CHB (right panel). Red circle indicates rs2981582, shown in bold black. c, Sequence conservation of 32 kb region in five species, relative to human sequence (http://pipeline.lbl.gov/methods.shtml). Red circle indicates rs2981582. SNPs in grey are those used in the initial tagging of known common HapMap SNPs within the block. SNPs in black are correlated with rs2981582 with r²>0.6 in European samples. Six SNPs in red were those consistent with being the causative variant on the basis of the genetic data (not excluded at odds of 100:1 relative to the SNP with the strongest association, rs7895676).