Although association with lung function at 4q24 is well established,5,7 we report two further independent signals of association at this locus (table 2; appendix pp 56–57, 99–101). The first (rs34712979) was localised to NPNT but independent of the previously reported signal, which spanned INTS12, GSTCD, and NPNT (appendix p 102). The second (rs2047409) was 552 kb from the signals at INTS12, GSTCD, and NPNT and was localised to TET2. The signal of association for rs34712979 was strongest in never smokers (p=9·62 × 10−16; table 2) and was weakly correlated (linkage disequilibrium r2=0·31) with another SNP in NPNT—rs6856422—which has also been identified as a novel secondary signal of association at 4q24 by an independent concurrent study of lung function in the general population.35 When rs6856422 was included as a covariate in our analysis, the signal for rs34712979 (p=9·62 × 10−16) was only slightly attenuated (p=4·66 × 10−11). The novel signal at TET2 (rs2047409) also showed strongest association in never smokers (p=1·31 × 10−8). rs2047409 is separated from the previously reported association of rs10516526 with FEV1 (GSTCD)5,7 by a recombination hotspot and
