Promoter landscape of maternal plasma DNA reveals predictive signatures of fetal growth restriction.

backgroundFetal growth restriction (FGR) causes serious pregnancy complications, but early detection is challenging. Promoter coverage patterns from maternal plasma cell-free DNA (cfDNA) offer a promising, non-invasive method for early FGR detection.

methodsThis retrospective multicenter study enrolled 788 singleton pregnancies (282 FGR and 506 matched controls) undergoing NIPT blood draw between 12 and 29 weeks of gestation from four hospitals in China. cfDNA promoter coverage was analyzed to identify transcription start site (TSS) features differentially covered in FGR. Feature selection was performed using LASSO regression. Three machine learning models—support vector machine (SVM), logistic regression (LR), and K-nearest neighbor (KNN)—were constructed. Model performance was assessed using area under the ROC curve (AUC), with an internal dataset used for validation. Functional enrichment analysis was conducted for biological interpretation.

resultsA total of 198 differentially covered TSS features were selected. In the validation dataset, both LR and SVM models demonstrated comparable predictive accuracy, with AUCs of 0.7 for LR and 0.69 for SVM. The associated genes were enriched in pathways related to placental and fetal development.

conclusionPromoter profiling of maternal plasma cfDNA represents a novel, non-invasive approach for mid-pregnancy FGR prediction. This method could enhance early detection of at-risk pregnancies and inform clinical decision-making in prenatal care.