Single-locus statistics for height associations of fifty-two SNPs from the thirty-four genomic regions investigated in this study.

1Only SNPs with an asterisk were genotyped in the CBR collection; the sample size for the remaining SNPs is 4,274. 2Meta-analysis P-values were calculated from study-specific best analysis P-values weighted by sample size. 3Betas and SE for the combined sample were calculated using inverse variance meta-analysis for height values normalised to z-scores. For family-based cohorts, such values were calculated in a subset of unrelated individuals (n = 1,381 for TwinsUK discovery and n = 1,403 for TwinsUK replication). 41 = Weedon et al. 2007; 2 = Lettre et al. 2008; 3 = Weedon et al. 2008, 4 = Sanna et al. 2008; 5 = Gudbjartsson et al. 2008. The SNPs having the lowest association P-value within in each region in the combined analysis are indicated in bold. For loci also described in the Weedon et al. study we report association P-values excluding the EPIC Norfolk, to account for the sample overlap with that study [9].