W5. DNA METHYLATION AS PROXY OF GENETIC AND PRENATAL PSYCHIATRIC RISK FACTORS

Background Genetic and prenatal risk factors contribute to neurodevelopmental conditions (NDCs). DNA methylation at birth may integrate their combined effects. Methods By leveraging data from two longitudinal birth cohorts, Generation R (Ntrain = 1856, Ntest = 476) and ALSPAC (Nvalidation= 847), we developed and validated methylation profile scores (MPSs) in cord blood, based on genetic and prenatal risk factors for NDCs. We assessed their individual and joint predictive performance for a broad range of childhood psychiatric outcomes measured repeatedly across childhood (ages 6-14). We additionally assessed if the MPSs could be combined into a composite “transmission load” MPS. Results We identified 11 genetic and prenatal risk factors for MPS construction. Two had previously validated MPSs for maternal smoking and gestational age, and we developed nine novel MPSs. Four of these novel MPSs were validated in the testing and external validation datasets: maternal age, birthweight, and polygenic scores (PGSs) for ADHD and schizophrenia. In the validation dataset, the correlations with their corresponding risk factors ranged from 0.29 for birthweight to 0.08 for maternal age. Additive models explained more variance than individual models. Across all psychiatric outcomes, the models that included all six MPSs explained, on average, 1.2% of the variance, and the model including both MPSs and their corresponding risk factors, explained additional 1% unique variance. The “transmission load” MPS did not replicate in the testing or the external validation dataset. Discussion We developed four novel MPSs, which may serve as valuable tools for future research. By integrating genetic and prenatal risk factors with DNAm at birth, our study offers new insights into their individual and additive effects on childhood psychiatric problems. Combining risk factors with epigenetics shows promise for improving early prediction of childhood psychiatric outcomes.