Limits...
DNA methylation mediates the effect of maternal smoking during pregnancy on birthweight of the offspring.

Küpers LK, Xu X, Jankipersadsing SA, Vaez A, la Bastide-van Gemert S, Scholtens S, Nolte IM, Richmond RC, Relton CL, Felix JF, Duijts L, van Meurs JB, Tiemeier H, Jaddoe VW, Wang X, Corpeleijn E, Snieder H - Int J Epidemiol (2015)

Bottom Line: We found 35 differentially methylated CpGs (FDR < 0.05) in newborns exposed vs unexposed to smoking, of which 23 survived Bonferroni correction (P < 1 × 10(-7)).These CpGs partly mediated the effect of maternal smoking on birthweight (Sobel P < 0.05) in meta-analysis of GECKO and the two replication cohorts.Functional network analysis suggested a role in activating the immune system.

View Article: PubMed Central - PubMed

Affiliation: Departments of Epidemiology, l.k.kupers@umcg.nl.

No MeSH data available.


Related in: MedlinePlus

Hypothetical mediation model explaining the variables in the mediation analysis. βa, effect estimate for smoking in the model: CpG = smoking + covariates. βb, effect estimate for CpG in the model: BW = CpG + covariates. βc, effect estimate for smoking in the model: BW = smoking + covariates. βc’: effect estimate for smoking in the model: BW = smoking + CpG + covariates.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4588868&req=5

dyv048-F1: Hypothetical mediation model explaining the variables in the mediation analysis. βa, effect estimate for smoking in the model: CpG = smoking + covariates. βb, effect estimate for CpG in the model: BW = CpG + covariates. βc, effect estimate for smoking in the model: BW = smoking + covariates. βc’: effect estimate for smoking in the model: BW = smoking + CpG + covariates.

Mentions: We tested the CpGs with FDR < 0.05 for mediation in the association between maternal smoking during pregnancy and birthweight, using the widely used method of Baron and Kenny31 and the Sobel test.32 As shown in Figure 1, mediation was considered to be present when: (i) smoking correlated with methylation level (βa); (ii) smoking correlated with birthweight without adjusting the model for the mediator (βc); (iii) differential methylation correlated with birthweight (βb); (iv) the association between smoking and birthweight decreased upon addition of methylation to the model (βc’); and (v) the Sobel test gave P < 0.05, indicating a decrease in the effect of smoking on birthweight after adjusting for the differentially methylated CpG. For those CpGs showing mediation, we tested the assumption that there is no interaction of the exposure and covariates with the mediator CpGs.33,34 For the mediating CpGs, we further calculated which part of the association between smoking and birthweight could be explained by the mediator using the formula:35(βa*βb)/((βa*βb)+βc′).The mediation effect βa*βb equals βc - βc’, thus this formula equals:βc-βc′/βc.Figure 1.


DNA methylation mediates the effect of maternal smoking during pregnancy on birthweight of the offspring.

Küpers LK, Xu X, Jankipersadsing SA, Vaez A, la Bastide-van Gemert S, Scholtens S, Nolte IM, Richmond RC, Relton CL, Felix JF, Duijts L, van Meurs JB, Tiemeier H, Jaddoe VW, Wang X, Corpeleijn E, Snieder H - Int J Epidemiol (2015)

Hypothetical mediation model explaining the variables in the mediation analysis. βa, effect estimate for smoking in the model: CpG = smoking + covariates. βb, effect estimate for CpG in the model: BW = CpG + covariates. βc, effect estimate for smoking in the model: BW = smoking + covariates. βc’: effect estimate for smoking in the model: BW = smoking + CpG + covariates.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4588868&req=5

dyv048-F1: Hypothetical mediation model explaining the variables in the mediation analysis. βa, effect estimate for smoking in the model: CpG = smoking + covariates. βb, effect estimate for CpG in the model: BW = CpG + covariates. βc, effect estimate for smoking in the model: BW = smoking + covariates. βc’: effect estimate for smoking in the model: BW = smoking + CpG + covariates.
Mentions: We tested the CpGs with FDR < 0.05 for mediation in the association between maternal smoking during pregnancy and birthweight, using the widely used method of Baron and Kenny31 and the Sobel test.32 As shown in Figure 1, mediation was considered to be present when: (i) smoking correlated with methylation level (βa); (ii) smoking correlated with birthweight without adjusting the model for the mediator (βc); (iii) differential methylation correlated with birthweight (βb); (iv) the association between smoking and birthweight decreased upon addition of methylation to the model (βc’); and (v) the Sobel test gave P < 0.05, indicating a decrease in the effect of smoking on birthweight after adjusting for the differentially methylated CpG. For those CpGs showing mediation, we tested the assumption that there is no interaction of the exposure and covariates with the mediator CpGs.33,34 For the mediating CpGs, we further calculated which part of the association between smoking and birthweight could be explained by the mediator using the formula:35(βa*βb)/((βa*βb)+βc′).The mediation effect βa*βb equals βc - βc’, thus this formula equals:βc-βc′/βc.Figure 1.

Bottom Line: We found 35 differentially methylated CpGs (FDR < 0.05) in newborns exposed vs unexposed to smoking, of which 23 survived Bonferroni correction (P < 1 × 10(-7)).These CpGs partly mediated the effect of maternal smoking on birthweight (Sobel P < 0.05) in meta-analysis of GECKO and the two replication cohorts.Functional network analysis suggested a role in activating the immune system.

View Article: PubMed Central - PubMed

Affiliation: Departments of Epidemiology, l.k.kupers@umcg.nl.

No MeSH data available.


Related in: MedlinePlus