The authors examined whether newborn cord blood DNA methylation profile scores (MPSs) can serve as molecular proxies for known genetic, prenatal and perinatal risk factors for neurodevelopmental conditions (NDCs) and related childhood psychiatric outcomes. The premise is that methylation measured at birth may capture downstream biological signatures of exposures and genetic liability, potentially improving understanding of how combined exposures relate to later psychiatric risk.
Analyses combined data from two prospective population-based birth cohorts. The authors developed cord blood MPSs corresponding to specific genetic and pre-/perinatal risk factors, and then evaluated the association of each MPS with its respective measured risk factor. They also examined the predictive performance of individual risk factors and MPSs, both separately and jointly, for a set of childhood psychiatric outcomes. Finally, they tested whether multiple MPSs could be aggregated into a composite metric termed the “transmission load” MPS.
The study used two longitudinal birth cohorts: Generation R and ALSPAC. Generation R provided a training set and a test set (N-train = 1,856; N-test = 476). ALSPAC served as an external validation cohort (N-validation = 832). Childhood psychiatric outcomes were assessed between ages 5 and 14 years and included eight outcomes (four broad and four specific), although the source text does not list the exact outcome names.
The investigators developed MPSs in cord blood for multiple predictors. They validated four novel MPSs: maternal age, birthweight, and genetic liability for ADHD and schizophrenia. They also included two previously validated MPSs: maternal smoking and gestational age. Reported correlations between MPSs and their corresponding measured risk factors were r = 0.08 to 0.29 for the four novel MPSs and r = 0.42 to 0.63 for the two previously validated MPSs, indicating replication of several methylation–exposure signals across cohorts.
The study evaluated how well risk factors, MPSs, and their combination explained variance in eight childhood psychiatric outcomes measured between ages 5 and 14. Modeling the six MPSs together with their corresponding measured risk factors explained on average 3.3% of variance in outcomes. By comparison, measured risk factors alone explained on average 1.8% of variance, and MPSs alone explained on average 1.6% of variance. These results indicate that MPSs and measured risk factors provide complementary information for explaining early psychiatric outcome variance.
The authors tested the feasibility of combining the individual MPSs into a composite “transmission load” MPS intended to capture aggregated prenatal/genetic signal. This composite did not replicate across cohorts. The lack of replication was attributed to heterogeneous contributions of the individual predictors across the two cohorts, meaning that the pattern of methylation–predictor relationships differed sufficiently between Generation R and ALSPAC to prevent stable aggregation into a single composite score.
Four novel cord blood MPSs (maternal age, birthweight, genetic liability for ADHD and schizophrenia) were validated alongside two previously reported MPSs (maternal smoking and gestational age). The combined use of MPSs with measured risk factors modestly increased the variance explained in childhood psychiatric outcomes compared with either type of predictor alone, suggesting these measures are complementary. The attempted composite “transmission load” MPS did not replicate, highlighting cohort heterogeneity in predictor contributions.
The authors propose that integrating genetic and prenatal risk factor information with cord blood DNA methylation may aid future research into early psychiatric risk and could help refine prediction models, although absolute variance explained in this study was small.
Data from Generation R are available upon reasonable request to the study director and subject to applicable rules and regulations. ALSPAC data are available on request under an approved proposal number; the source provides links for data access procedures. Ethical approvals were obtained from the Medical Ethical Committee of Erasmus MC for Generation R and from the ALSPAC Ethics and Law Committee and Local Research Ethics Committees for ALSPAC. Participant consent procedures were followed as described.
Funding sources reported include HorizonEurope Research and Innovation Programme (FAMILY, HappyMums), the European Research Council (TEMPO), Erasmus MC, Netherlands Organization for Health Research and Development, National Institute of Child and Human Development, UK Medical Research Council and Wellcome, University of Bristol, and others. The authors declared no competing interests.
The source article reports cohort sizes, the set of MPSs developed and validated, correlation ranges between MPSs and predictors, variance explained by predictors and MPSs, and the lack of replication for the composite MPS. The source did not provide the full list of the eight specific psychiatric outcome names in the abstract, nor detailed effect estimates for each outcome in this summary; those details would be found in the full preprint and supplementary material referenced by the authors.