Detection of early-onset severe preeclampsia by cell-free DNA fragmentome

Abstract

Early-onset severe preeclampsia (EO-PE) is a distinct and highly consequential form of preeclampsia (PE), presenting significant challenges for early detection. Here, we investigated the fragmentation pattern of plasma cell-free DNA (cfDNA) in EO-PE patients. We uncovered that the nucleotide composition at the 5 end (i.e. ends motif) of plasma cfDNA showed a unique pathological preference in EO-PE pregnancies and gestational-psychology preference in healthy pregnancies. By integrating 91 EO-PE specific motifs into a machine-learning model, we achieved accurate prediction of EO-PE development in early pregnancies. Remarkably, our model demonstrated robust performance in an independent cohort of 74 early pregnancies and 1,241 non-invasive prenatal testing (NIPT) samples with ultra-low sequencing depth. Additionally, we revealed that these PE-specific motif signatures lacked tissue specificity, originating extracellularly, and were associated with the abnormal concentration of DNA fragmentation factor subunit beta (DFFB) in EO-PE patients plasma. These findings establish the plasma DNA fragmentome as a non-invasive and cost-effective biomarker that can be simultaneously captured during NIPT for early EO-PE detection and provide valuable insights into cfDNA production mechanisms in preeclampsia patients.