Direct and indirect genetic effects of birthweight predisposition on child DNA methylation at birth

Birthweight is heritable and strongly associated with epigenetic differences at birth. It is unclear whether a genetic birthweight score is associated with DNA methylation (DNAm) and, if so, whether through direct genetic effects (i.e. child genotype) or indirect genetic effects (i.e. parental genotype, independent of child genotype), which are suspected to be mediated by the prenatal environment (e.g. metabolic factors). We constructed polygenic scores (PGS) for birthweight predisposition in mothers and children using an existing 'pre-adjusted' GWAS assessing maternal indirect effects (maternal genotype on offspring birthweight, correcting for offspring genotype) or fetal effects (offspring genotype on offspring birthweight, correcting for maternal genotype) in 2116 families from the Generation R Study. Offspring DNAm levels at 829 birthweight-related sites were then regressed on both maternal and fetal effect PGS. Additionally, we aggregated 829 DNAm sites into a methylation profile score for birthweight (MPS-BW) and tested which pregnancy health factors (n = 13) might mediate genetic effects. We identified six DNAm sites associated with maternal indirect effects and the birthweight MPS revealed indirect genetic associations, as well. Paternal genotype was not associated with offspring DNAm in this study but was limited by lower sample size and lack of GWAS on paternal indirect effects, thus requiring a posteriori disentanglement of direct and indirect effects. Gestational age partly mediated maternal indirect effects on DNAm, but no other mediators were identified. Results did not depend on offspring sex. This study presents first evidence of maternal genetic effects on offspring birthweight being associated with offspring epigenetic patterns at birth.