Machine learning–driven decoding of maternal immune signatures in repeated pregnancy loss
4 affiliations
- 1Center for Biomedical Computing, Korea Institute of Science and Technology Information (KISTI), andDepartment of Data and HPC Science, Korea National University of Science and Technology (UST), Daejeon, Republic of Korea
- 2Department of Bio and Brain Engineering, Korea Advanced Institute of Science and Technology, Daejeon, Republic of Korea, and Cambridge Stem Cell Institute, University of Cambridge, Cambridge, United Kingdom
- 3Department of Obstetrics and Gynecology, College of Medicine, Myunggok Medical Research Institute, Konyang University, Daejeon, Republic of Korea
- 4Department of Biological Sciences, Sungkyunkwan University, Suwon, Republic of Korea
Abstract
Repeated pregnancy loss (RPL) is a multifactorial condition in which the underlying immunological mechanisms, particularly the disruption of maternal-fetal tolerance, remain incompletely understood. Although immune tolerance is critical for pregnancy success, the specific immune dysregulations contributing to RPL, particularly in euploid pregnancies, have been difficult to characterize. To address this, we performed single-cell RNA sequencing of decidual tissues from RPL patients and first-trimester controls. Our analysis initially revealed elevated expression of a transcriptional module of immune activation genes in RPL decidual tissues. To dissect the cellular drivers of this complex landscape, we employed genotype-based origin analysis coupled with a supervised machine learning model and a transformer-based foundation model (scGPT). This hierarchical approach prioritized maternal T cells over other immune subsets as the population carrying the most distinct and generalizable RPL-associated signatures. Through the convergence of computational drug repurposing, network centrality analysis, and a rigorous origin-controlled expression filtering strategy, we identified CXCR4 and JUN as druggable molecular candidates strongly associated with this T-cell dysregulation. Collectively, our machine learning–driven approach characterizes the maternal immune landscape of euploid RPL in the context of immune tolerance breakdown, and nominates candidate targets for future functional investigation. Author summary: Repeated Pregnancy Loss (RPL) is a devastating condition affecting many couples. It is particularly perplexing when it occurs even though the fetus is genetically normal, leaving doctors and patients without clear answers. Successful pregnancy requires the mother’s immune system to tolerate the fetus. However, the exact reasons why this tolerance fails in RPL have remained a mystery. In this study, we first found that in these unexplained miscarriages, maternal immune cells in the decidual tissue show increased activity of genes involved in immune activation. To characterize this immune activation in more detail, we applied advanced artificial intelligence to analyze the gene activity of individual cells at the interface between the mother and the fetus. This analysis identified maternal T cells as the immune cells most strongly and consistently linked to RPL. Furthermore, by mapping the complex molecular networks governing these cells, we identified specific molecules (CXCR4 and JUN) strongly associated with this immune dysregulation. This study provides a new way of understanding RPL in relation to immune tolerance failure and identifies these specific molecules as candidate targets for developing future treatments.
Related
- Disentangling maternal pregnancy and fetal-specific plasma signatures through integrated proteomics and site-specific N-glycoproteomics
- Pregnancy Impacts Maternal Sleep: Focus on Kynurenine Pathway Activation in Female Rats
- Maternal gut microbiota, tryptophan metabolism and immune homeostasis: Implications for the prevention of adverse pregnancy complications
- Neonatal T cell maternal microchimerism is common and associates with early pregnancy SARS2 vaccination
- Maternal Metabolic Profiles, Their Relationships with Progesterone and Pregnancy-Associated Glycoprotein (PAG), and Biomarker Potential During Early Pregnancy in Lactating Ewes.
BibTeX
@article{ko2026machine,
title = {{Machine learning–driven decoding of maternal immune signatures in repeated pregnancy loss}},
author = {Tae Lyun Ko and Jaesub Park and Dongju Leem and Junho Kim and Jae won Han and Jin Sol Park and Sung Ki Lee and Hyojung Paik},
journal = {PLOS Computational Biology},
year = {2026},
month = sep,
doi = {10.1371/journal.pcbi.1014806},
url = {https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1014806},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Machine learning–driven decoding of maternal immune signatures in repeated pregnancy loss
AU - Tae Lyun Ko
AU - Jaesub Park
AU - Dongju Leem
AU - Junho Kim
AU - Jae won Han
AU - Jin Sol Park
AU - Sung Ki Lee
AU - Hyojung Paik
PY - 2026
DA - 2026/09/17
JO - PLOS Computational Biology
DO - 10.1371/journal.pcbi.1014806
UR - https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1014806
AB - Repeated pregnancy loss (RPL) is a multifactorial condition in which the underlying immunological mechanisms, particularly the disruption of maternal-fetal tolerance, remain incompletely understood. Although immune tolerance is critical for pregnancy success, the specific immune dysregulations contributing to RPL, particularly in euploid pregnancies, have been difficult to characterize. To address this, we performed single-cell RNA sequencing of decidual tissues from RPL patients and first-trimester controls. Our analysis initially revealed elevated expression of a transcriptional module of immune activation genes in RPL decidual tissues. To dissect the cellular drivers of this complex landscape, we employed genotype-based origin analysis coupled with a supervised machine learning model and a transformer-based foundation model (scGPT). This hierarchical approach prioritized maternal T cells over other immune subsets as the population carrying the most distinct and generalizable RPL-associated signatures. Through the convergence of computational drug repurposing, network centrality analysis, and a rigorous origin-controlled expression filtering strategy, we identified CXCR4 and JUN as druggable molecular candidates strongly associated with this T-cell dysregulation. Collectively, our machine learning–driven approach characterizes the maternal immune landscape of euploid RPL in the context of immune tolerance breakdown, and nominates candidate targets for future functional investigation. Author summary: Repeated Pregnancy Loss (RPL) is a devastating condition affecting many couples. It is particularly perplexing when it occurs even though the fetus is genetically normal, leaving doctors and patients without clear answers. Successful pregnancy requires the mother’s immune system to tolerate the fetus. However, the exact reasons why this tolerance fails in RPL have remained a mystery. In this study, we first found that in these unexplained miscarriages, maternal immune cells in the decidual tissue show increased activity of genes involved in immune activation. To characterize this immune activation in more detail, we applied advanced artificial intelligence to analyze the gene activity of individual cells at the interface between the mother and the fetus. This analysis identified maternal T cells as the immune cells most strongly and consistently linked to RPL. Furthermore, by mapping the complex molecular networks governing these cells, we identified specific molecules (CXCR4 and JUN) strongly associated with this immune dysregulation. This study provides a new way of understanding RPL in relation to immune tolerance failure and identifies these specific molecules as candidate targets for developing future treatments.
SN - 1553-7358
ER -