Early Pregnancy DNA Methylation Signatures as Predictors of Antenatal Depressive Symptoms: A longitudinal study of DNA methylation changes

Background Antenatal depressive symptoms (ADS) are common and underdiagnosed, particularly in low- and middle-income countries, and are associated with adverse maternal and offspring outcomes. Current screening relies on subjective symptom reporting, limiting early identification and prevention. Epigenetic modifications, particularly DNA methylation, offer a promising avenue for objective, early biomarkers of depression risk during pregnancy. Methods In this nested case–control design within the STRiDE (Stratification of Risk of Diabetes in Early Pregnancy study) prospective cohort, 189 pregnant women with no depressive symptoms in early pregnancy (<16 weeks of gestation. PHQ-9 ≤4) were followed longitudinally. 89 ADS individuals were identified by the emergence of depressive symptoms at 24–28 weeks of gestation (PHQ-9 ≥5), while 100 women remained symptom-free (Controls). Genome-wide DNA methylation profiling of early-pregnancy peripheral blood was performed using the Illumina EPIC 850K array. Epigenome-wide association analyses were combined with machine-learning approaches to identify predictive CpG panels. Model robustness was assessed using bootstrap validation, and a methylation risk score (MRS) was constructed. Functional enrichment analyses were conducted to explore biological pathways. Results Epigenome-wide analysis identified 2,447 differentially methylated positions associated with subsequent ADS. A robust panel of 10 CpGs in early pregnancy predicted later ADS with excellent performance (testing AUC=0.99. bootstrap-validated AUC=0.91), independent of maternal risk factors. The MRS markedly outperformed traditional clinical predictors (AUC=0.94) and further improved prediction when combined with maternal characteristics (AUC=0.95). ADS-associated methylation changes were enriched in neurodevelopmental, synaptic, immune, and metabolic signaling pathways. Limited concordance with placental methylation suggested maternal-specific epigenetic regulation. Conclusions Early-pregnancy DNA methylation signatures can predict antenatal depressive symptoms before clinical onset. This blood-based 10-CpG biomarker panel offers a biologically informed and objective tool for early risk stratification, with the potential to enable preventive interventions and enhance perinatal mental health care, particularly in resource-constrained settings.