are presented in Supplementary Fig. 6. Meta-analysis of the 3 analyses successfully recovered 5 out of 7 loci reported as COPD-associated (Fig. 4 and Table 2) at genome-wide significance (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$P\, < \, 5\times {10}^{-8}$$\end{document}P<5×10−8, Supplementary Table 2). The direction of association for all recovered SNPs was the same as previously reported28. The two SNPs that did not exceed the genome-wide significance threshold were nominally associated at P < 0.05 in our analysis. These two SNPs were missing in Analysis 1 (COPD cases with WGS data from COPDGene EA Vs. smoking controls with array data from COGEND EA) due to the filters applied with the protocol; the reduced power caused by their missingness likely explain the lower significance level observed.Table 2Recovery of GWAS-identified variants, following application of our protocol to each of 3 GWAS and their meta-analysis, compared to published risk loci for COPD with combined data from COPDGene, ECLIPSE, NETT/NAS, and GenKOLS (Norway)24.SNPPositionRisk AlleleRelated geneReported (N = 12,337)Current meta-analysis (N = 10,461)ORP-valueORDirectionP-valuers12914385chr15:78898723TCHRNA31.362.70E-161.28+++3.35E-16rs4416442chr4:89866713CFAM13A1.369.44E-151.21+++2.66E-10rs793727chr19:41302706CCYP2A60.742.88E-090.84---1.91E-08rs4846480chr1:218598469ATGFB21.261.25E-071.19+++9.37E-08rs13141641chr4:145506456THHIP1.393.66E-151.23?++*2.64E-07rs754388chr14:93115410CRIN31.336.69E-081.12?++*0.020rs626750chr11:102720945GMMP3/121.365.35E-091.14?++*0.005*The question mark “?” means the SNP is missing from the first analysis, and it may result in reduced power in the final meta-analysis.
