“…We were initially surprised that this large-effect signal had not been reported by larger IPF GWASs, such as that of Allen and colleagues ( 2 ) or Zhou and colleagues ( 3 ). However, we noted that previous association testing of rs367849850, rs3930589, and other variants with which they are in strong linkage disequilibrium was inconsistent because of poor regional coverage after array genotyping and imputation ( Figure 1D ); additionally, using biobank-derived IPF cases may attenuate association signals ( 1 , 8 ). Nevertheless, we observed replication in data from two independent studies: the Global Biobank Meta-analysis Initiative (rs367849850: OR, 1.3; P = 6.5 × 10 −4 ; rs3930589: OR, 1.2; P = 7.0 × 10 −4 ) ( 5 ) and the discovery phase of a published IPF GWAS with clinically derived cases (rs367849850: OR, 2.4; P = 1.8 × 10 −6 ) ( 9 ).…”