Fig. 6: Comparison of the size of effects and P values with and without adjustment for parental methylation.

The effect sizes and observed −log₁₀ P values (calculated from t statistics) were derived from two-sided t tests for comparing the DNA methylation levels between ART conceived newborns (n = 962) and naturally conceived newborns (n = 983). a The effect sizes on the x-axis were adjusted for maternal age, maternal smoking, maternal BMI before pregnancy, parity, child sex, and plate number, and the effect sizes on the y-axis were adjusted for all the covariates listed and parental methylation. The Pearson’s correlation coefficient was 0.95. b The observed −log₁₀ P values were calculated from the effect sizes in panel (a) and their standard errors. Blue dots indicate the significant CpG sites at FDR < 0.01, while gray dots indicate nonsignificant CpG sites. c The expected −log₁₀ P values adjusted for maternal age, maternal smoking, maternal BMI before pregnancy, parity, child sex, and plate number are on the x-axis, and the observed P values are on the y-axis. Blue dots indicate the significant CpG sites at FDR < 0.01, while gray dots indicate nonsignificant CpG sites. The slope in gray indicates the null distribution of no differences between ART and naturally conceived. d The expected −log₁₀ P values adjusted for maternal age, maternal smoking, maternal BMI before pregnancy, parity, child sex, plate number, and parental DNA methylation are on the x-axis, and the observed P values are on the y-axis. Blue dots indicate the significant CpG sites at FDR < 0.01, while gray dots indicate nonsignificant CpG sites. The slope in gray indicates the null distribution of no differences between ART and naturally conceived. Source data are provided as a Source Data file.