Table 3.
Total physical activity interactions with a SNP near the ERC1/WNT5B locus and TBLH-BMC
| GG | AG | AA | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| SNP | Nearest Gene | Risk Allelea | Sex | Site | Cohort | P-interaction | Estimate (95% CI) | P-value | Estimate (95% CI) | P-value | Estimate (95% CI) | P-value |
| rs2887571 | ERC1/WNT5B | A | M | TB | Discovery (N=444) | 7.1×10−5 | −0.02 (−0.04, 0.002) | 0.069 | 0.01 (−0.005, 0.02) | 0.212 | 0.04 (0.02, 0.05) | 1.2×10−6 |
| Replication (N=241) | 0.999 | 0.02 (−0.22, 0.26) | 0.867 | 0.02 (−0.10, 0.14) | 0.734 | 0.02 (−0.10, 0.14) | 0.734 | |||||
| Both (N=685) | 6.0×10−5 | −0.02 (−0.04, 0.001) | 0.066 | 0.01 (−0.004, 0.02) | 0.204 | 0.04 (0.02, 0.05) | 8.1×10−7 | |||||
| rs2887571 | ERC1/WNT5B | A | F | TB | Discovery (N=485) | 0.027 | −0.03 (−0.08, 0.01) | 0.094 | −0.01 (−0.03, 0.01) | 0.409 | 0.02 (0.002, 0.03) | 0.029 |
| Replication (N=244) | 0.463 | −0.08 (−0.2, 0.09) | 0.505 | −0.02 (−0.10, 0.06) | 0.689 | 0.04 (−0.06, 0.14) | 0.571 | |||||
| Both (N=729) | 0.031 | −0.03 (−0.07, 0.01) | 0.104 | −0.01 (−0.03, 0.01) | 0.434 | 0.02 (0.002, 0.03) | 0.030 | |||||
| rs2887571 | ERC1/WNT5B | A | M+F | TB | Discovery (N=816) | 6.7×10−4 | −0.02 (−0.04, 0.004) | 0.109 | 0.01 (−0.005, 0.02) | 0.296 | 0.03 (0.02, 0.04) | 8.8×10−6 |
| Replication (N=484) | 0.853 | −0.003 (−0.17, 0.16) | 0.975 | 0.01 (−0.07, 0.09) | 0.858 | 0.02 (−0.06, 0.09) | 0.691 | |||||
| Both (N=1300) | 9.1×10−4 | −0.01 (−0.03, 0.004) | 0.132 | 0.01 (−0.004, 0.02) | 0.255 | 0.03 (0.01, 0.04) | 8.2×10−6 | |||||
The estimates represent the change in bone Z-scores per additional hour per day spent physically active, by rs2887571 genotype.
Risk allele was associated with lower bone mineral density in a genome-wide association study of adults(20).
Discovery: Analyses involving participants in the longitudinal discovery cohort using linear mixed models; Replication: Analyses involving participants in the cross-sectional replication cohort using linear regression; Both: Analyses involving participants in the discovery and replication cohorts using linear mixed models; all models were adjusted for age, puberty stage, body mass index Z-score, and dietary calcium intake (gender was included as a covariate in the male and females combined analyses; older sibling was removed for the male and female combined analyses). P-interactions in bold survive correction for multiple testing for 67 independent loci. Abbreviations: F, female; M, male; SNP, single nucleotide polymorphism; TB, total body less head bone mineral content.