Xia Xu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Xia Xu has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Jianying Yan
College of Clinical Medicine for Obstetrics & Gynecology and Pediatrics, Fujian Medical University Fujian Maternity and Child Health Hospital
8 shared publications Zewei Liang
College of Clinical Medicine for Obstetrics and Gynecology and Pediatrics, Fujian Medical University Fujian Maternity and Child Health Hospital
5 shared publicationsChengcheng Jin
College of Clinical Medicine for Obstetrics and Gynecology and Pediatrics, Fujian Medical University Fujian Maternity and Child Health Hospital
4 shared publicationsXingyi Qi
College of Clinical Medicine for Obstetrics and Gynecology and Pediatrics, Fujian Medical University Fujian Maternity and Child Health Hospital
4 shared publications- Jianying Yan
College of Clinical Medicine for Obstetrics and Gynecology and Pediatrics, Fujian Medical University Fujian Maternity and Child Health Hospital
3 shared publications Xinying Liu
College of Clinical Medicine for Obstetrics and Gynecology and Pediatrics, Fujian Medical University Fujian Maternity and Child Health Hospital
3 shared publications
Latest Publications
- Open Access
- 22 October 2026
- Published by Frontiers Media SA in Frontiers in Pharmacology
- Vol. 17, 1928244
Quercetin ameliorates gestational diabetes mellitus-induced placental dysfunction by modulating the AGE-RAGE-associated oxidative and inflammatory signaling network
Background: Gestational diabetes mellitus (GDM) induces placental structural and functional disturbances associated with dysregulated AGE-RAGE/ROS signaling and inflammatory responses. Existing flavonoid-related studies merely observed changes in pathway protein levels, lacking direct experimental proof of molecular binding between candidate compounds and core receptors. This study aimed to investigate the potential involvement of RAGE in the protective effects of quercetin and to characterize the associated AGE-RAGE/ROS/NF-κB-related signaling changes in GDM placental injury. Methods: Network pharmacology was used to screen quercetin’s therapeutic targets for GDM. Molecular docking was performed using the human RAGE structure to predict the potential interaction between quercetin and RAGE, and molecular dynamics simulation was conducted to evaluate the stability of the predicted complex. A combined streptozotocin-high-fat diet GDM rat model and 30 mM high-glucose HTR-8/SVneo trophoblast injury model were constructed. FPS-ZM1 was used as a pharmacological tool to perturb RAGE-associated signaling in vitro . ROS, inflammatory cytokines, and key pathway gene and protein expression were examined to characterize the associated signaling changes. Results: Network pharmacology identified SERPINE1 (a critical regulator of placental vascular remodeling) as the core hub gene, with the AGE-RAGE pathway showing the highest enrichment score. Molecular docking and molecular dynamics simulations predicted a dynamically stable complex formed between quercetin and the ligand-binding domain of RAGE. GDM induction was associated with increased AGE accumulation, RAGE expression, ROS production, phosphorylated NF-κB p65, and SERPINE1 expression, together with an increase in Nrf2 expression, whereas quercetin treatment attenuated these alterations. Quercetin intervention also improved glucose tolerance and placental pathological changes. Conclusion: Quercetin alleviates GDM-associated placental dysfunction by improving metabolic abnormalities, oxidative imbalance, and inflammatory responses. Computational analysis suggests a potential interaction between quercetin and RAGE. These findings provide evidence that quercetin may regulate the AGE-RAGE-associated oxidative and inflammatory signaling network and improve placental homeostasis under diabetic conditions.
- 22 October 2026
- Published by Frontiers Media SA in Frontiers in Pharmacology
- Vol. 17, 1928244
Open AccessQuercetin ameliorates gestational diabetes mellitus-induced placental dysfunction by modulating the AGE-RAGE-associated oxidative and inflammatory signaling network
Background: Gestational diabetes mellitus (GDM) induces placental structural and functional disturbances associated with dysregulated AGE-RAGE/ROS signaling and inflammatory responses. Existing flavonoid-related studies merely observed changes in pathway protein levels, lacking direct experimental proof of molecular binding between candidate compounds and core receptors. This study aimed to investigate the potential involvement of RAGE in the protective effects of quercetin and to characterize the associated AGE-RAGE/ROS/NF-κB-related signaling changes in GDM placental injury. Methods: Network pharmacology was used to screen quercetin’s therapeutic targets for GDM. Molecular docking was performed using the human RAGE structure to predict the potential interaction between quercetin and RAGE, and molecular dynamics simulation was conducted to evaluate the stability of the predicted complex. A combined streptozotocin-high-fat diet GDM rat model and 30 mM high-glucose HTR-8/SVneo trophoblast injury model were constructed. FPS-ZM1 was used as a pharmacological tool to perturb RAGE-associated signaling in vitro . ROS, inflammatory cytokines, and key pathway gene and protein expression were examined to characterize the associated signaling changes. Results: Network pharmacology identified SERPINE1 (a critical regulator of placental vascular remodeling) as the core hub gene, with the AGE-RAGE pathway showing the highest enrichment score. Molecular docking and molecular dynamics simulations predicted a dynamically stable complex formed between quercetin and the ligand-binding domain of RAGE. GDM induction was associated with increased AGE accumulation, RAGE expression, ROS production, phosphorylated NF-κB p65, and SERPINE1 expression, together with an increase in Nrf2 expression, whereas quercetin treatment attenuated these alterations. Quercetin intervention also improved glucose tolerance and placental pathological changes. Conclusion: Quercetin alleviates GDM-associated placental dysfunction by improving metabolic abnormalities, oxidative imbalance, and inflammatory responses. Computational analysis suggests a potential interaction between quercetin and RAGE. These findings provide evidence that quercetin may regulate the AGE-RAGE-associated oxidative and inflammatory signaling network and improve placental homeostasis under diabetic conditions.
- Open Access
- 3 July 2026
- Published by Wiley in The Journal of Clinical Hypertension
- Vol. 28 (7) , e70290
Impact of Maternal Risk Stratification on Clinical Characteristics and Perinatal Outcomes in Singleton Pregnancies With Preeclampsia
This retrospective study of 4143 women at Fujian Maternal and Child Health Hospital (2012–2022) investigated how maternal risk levels influence clinical characteristics and perinatal outcomes in singleton pregnancies complicated by preeclampsia. Utilizing class‐imbalance‐adjusted and random forest algorithms, participants were stratified into high‐risk (17.38%), moderate‐risk (34.08%), and low‐risk (48.54%) groups. The analysis revealed that women in the high‐risk category were diagnosed at a significantly earlier gestational age and faced higher rates of severe, early‐onset, and preterm preeclampsia compared to the other groups. Although high‐risk pregnancies were associated with increased incidences of renal impairment, hypoproteinemia, and aminotransferase abnormalities, no significant differences were observed in rates of thrombocytopenia, placental abruption, postpartum hemorrhage, or stillbirth. Regarding delivery and neonatal outcomes, the high‐risk group demonstrated markedly higher rates of Cesarean sections, fetal distress, and neonatal intensive care unit (NICU) admissions. Key drivers of these adverse outcomes included chronic hypertension, chronic kidney disease, prior fetal growth restriction (FGR), and maternal age over 40 years. Ultimately, while high‐risk status is a strong predictor of early‐onset disease and specific maternal‐neonatal complications, it does not uniformly predict all adverse fetal outcomes, suggesting that clinical management should focus heavily on the specific impacts of hypertension, renal health, and advanced maternal age.
- 3 July 2026
- Published by Wiley in The Journal of Clinical Hypertension
- Vol. 28 (7) , e70290
Open AccessImpact of Maternal Risk Stratification on Clinical Characteristics and Perinatal Outcomes in Singleton Pregnancies With Preeclampsia
This retrospective study of 4143 women at Fujian Maternal and Child Health Hospital (2012–2022) investigated how maternal risk levels influence clinical characteristics and perinatal outcomes in singleton pregnancies complicated by preeclampsia. Utilizing class‐imbalance‐adjusted and random forest algorithms, participants were stratified into high‐risk (17.38%), moderate‐risk (34.08%), and low‐risk (48.54%) groups. The analysis revealed that women in the high‐risk category were diagnosed at a significantly earlier gestational age and faced higher rates of severe, early‐onset, and preterm preeclampsia compared to the other groups. Although high‐risk pregnancies were associated with increased incidences of renal impairment, hypoproteinemia, and aminotransferase abnormalities, no significant differences were observed in rates of thrombocytopenia, placental abruption, postpartum hemorrhage, or stillbirth. Regarding delivery and neonatal outcomes, the high‐risk group demonstrated markedly higher rates of Cesarean sections, fetal distress, and neonatal intensive care unit (NICU) admissions. Key drivers of these adverse outcomes included chronic hypertension, chronic kidney disease, prior fetal growth restriction (FGR), and maternal age over 40 years. Ultimately, while high‐risk status is a strong predictor of early‐onset disease and specific maternal‐neonatal complications, it does not uniformly predict all adverse fetal outcomes, suggesting that clinical management should focus heavily on the specific impacts of hypertension, renal health, and advanced maternal age.
- Open Access
- 6 March 2026
- Published by Springer Nature in BMC Medical Informatics and Decision Making
- Vol. 26 (1)
Phenotypic subclassification of preeclampsia through cluster analysis of preterm birth-related factors
Preeclampsia is a multifactorial hypertensive disorder of pregnancy with heterogeneous clinical presentation and outcomes. Current classification systems inadequately capture the substantial clinical heterogeneity of preeclampsia. This study aimed to identify and characterize distinct phenotypic clusters of the disease using preterm birth-related risk factors. This retrospective cohort study included 4,132 singleton pregnancies with preeclampsia between January 2014 and December 2024. We performed unsupervised k-means clustering based on 22 preterm birth-related clinical and biochemical variables. Maternal characteristics and pregnancy outcomes were compared across the identified clusters. Random forest algorithm was employed to evaluate feature importance. Five distinct clinical clusters were identified. Cluster 1 (N = 740) exhibited metabolic-inflammatory features. Cluster 2 (N = 1090) was characterized by multiple comorbidities, including pre-gestational diabetes (39.2%) and fetal growth restriction (47.8%). Cluster 3 (N = 732) featured a heightened systemic inflammatory response. Cluster 4 (N = 707) presented with the most benign clinical features. Cluster 5 (N = 863) was distinguished by a high prevalence of uterine mechanical factors. These clusters demonstrated significant differences in preterm birth rates and other adverse pregnancy outcomes (P < 0.05). Random forest analysis confirmed the discriminative power of the selected variables. Cluster analysis reveals five clinically distinct preeclampsia phenotypes with direct implications for preterm birth risk. This refined classification provides a foundation for investigating distinct pathophysiological mechanisms and supports the development of individualized management strategies.
- 6 March 2026
- Published by Springer Nature in BMC Medical Informatics and Decision Making
- Vol. 26 (1)
Open AccessPhenotypic subclassification of preeclampsia through cluster analysis of preterm birth-related factors
Preeclampsia is a multifactorial hypertensive disorder of pregnancy with heterogeneous clinical presentation and outcomes. Current classification systems inadequately capture the substantial clinical heterogeneity of preeclampsia. This study aimed to identify and characterize distinct phenotypic clusters of the disease using preterm birth-related risk factors. This retrospective cohort study included 4,132 singleton pregnancies with preeclampsia between January 2014 and December 2024. We performed unsupervised k-means clustering based on 22 preterm birth-related clinical and biochemical variables. Maternal characteristics and pregnancy outcomes were compared across the identified clusters. Random forest algorithm was employed to evaluate feature importance. Five distinct clinical clusters were identified. Cluster 1 (N = 740) exhibited metabolic-inflammatory features. Cluster 2 (N = 1090) was characterized by multiple comorbidities, including pre-gestational diabetes (39.2%) and fetal growth restriction (47.8%). Cluster 3 (N = 732) featured a heightened systemic inflammatory response. Cluster 4 (N = 707) presented with the most benign clinical features. Cluster 5 (N = 863) was distinguished by a high prevalence of uterine mechanical factors. These clusters demonstrated significant differences in preterm birth rates and other adverse pregnancy outcomes (P < 0.05). Random forest analysis confirmed the discriminative power of the selected variables. Cluster analysis reveals five clinically distinct preeclampsia phenotypes with direct implications for preterm birth risk. This refined classification provides a foundation for investigating distinct pathophysiological mechanisms and supports the development of individualized management strategies.
- Open Access
- 9 January 2026
- Published by Wolters Kluwer Health in Medicine
- Vol. 105 (2) , e46002
Immune cell signatures and preeclampsia: Unveiling causal links through genome-wide association studies
The immune-placental interaction could potentially contribute both causally and therapeutically to preeclampsia (PE). However, previous observational studies have not demonstrated a causal association between these 2 factors. Exploring the causal relationship between immune cells and PE provides genetic evidence that immune cells play a role in PE risk. Using a large dataset of genome-wide association studies, we conducted a bidirectional Mendelian randomization (MR) analysis to determine whether 731 immune cell signatures were correlated with PE, which included 4 types of immune signatures. After false discovery rate (FDR) correction, 3 immunophenotypes were identified to be significantly associated with PE risk: human leukocyte antigen on dendritic cells (odds ratio [OR] = 1.072, 95% confidence interval [CI] = 1.0352–1.110, P FDR = .012), absolute T cell count (OR = 1.477, 95% CI = 1.281–1.703, P FDR = 1.71 × 10 −5 ), and absolute lymphocyte count (OR = 1.514, 95% CI = 1.306–1.755, P FDR = 1.58 × 10 −5 ). Furthermore, PE did not affect immunophenotypes significantly after FDR adjustment, but 12 suggestive immunophenotypes were detected without FDR adjustment, including CD3 on activated and secreting CD4 regulatory T cell ( β = 0.467, P = .039), CD3 on CD39 + CD4 + T cell ( β = 0.508, P = .024), CD3 on CD45RA + CD4 + T cell ( β = 0.515, P = .022), CD3 on CD28 + CD45RA + CD8 + T cell ( β = 0.494, P = .028), CD3 on CD39 + CD8 + T cell ( β = 0.486, P = .029), CD45 on HLA DR + CD8 + T cell ( β = 0.518, P = .018), CD3 on secreting CD4 regulatory T cell ( β = 0.508, P = .025), CD25 on resting CD4 regulatory T cell ( β = 0.496, P = .018), CD4 on CD39 + activated CD4 regulatory T cell ( β = 0.445, P = .048), CD3 on CD28 + CD45RA − CD8 + T cell ( β = 0.498, P = .024), CD3 on effector memory CD4 + T cell ( β = 0.614, P = .007), and CD3 on activated CD4 regulatory T cell( β = 0.444, P = .049). Using genetic methods, this study uncovers the close relationship between immune cells and PE, which provides guidance for future clinical trials.
- 9 January 2026
- Published by Wolters Kluwer Health in Medicine
- Vol. 105 (2) , e46002
Open AccessImmune cell signatures and preeclampsia: Unveiling causal links through genome-wide association studies
The immune-placental interaction could potentially contribute both causally and therapeutically to preeclampsia (PE). However, previous observational studies have not demonstrated a causal association between these 2 factors. Exploring the causal relationship between immune cells and PE provides genetic evidence that immune cells play a role in PE risk. Using a large dataset of genome-wide association studies, we conducted a bidirectional Mendelian randomization (MR) analysis to determine whether 731 immune cell signatures were correlated with PE, which included 4 types of immune signatures. After false discovery rate (FDR) correction, 3 immunophenotypes were identified to be significantly associated with PE risk: human leukocyte antigen on dendritic cells (odds ratio [OR] = 1.072, 95% confidence interval [CI] = 1.0352–1.110, P FDR = .012), absolute T cell count (OR = 1.477, 95% CI = 1.281–1.703, P FDR = 1.71 × 10 −5 ), and absolute lymphocyte count (OR = 1.514, 95% CI = 1.306–1.755, P FDR = 1.58 × 10 −5 ). Furthermore, PE did not affect immunophenotypes significantly after FDR adjustment, but 12 suggestive immunophenotypes were detected without FDR adjustment, including CD3 on activated and secreting CD4 regulatory T cell ( β = 0.467, P = .039), CD3 on CD39 + CD4 + T cell ( β = 0.508, P = .024), CD3 on CD45RA + CD4 + T cell ( β = 0.515, P = .022), CD3 on CD28 + CD45RA + CD8 + T cell ( β = 0.494, P = .028), CD3 on CD39 + CD8 + T cell ( β = 0.486, P = .029), CD45 on HLA DR + CD8 + T cell ( β = 0.518, P = .018), CD3 on secreting CD4 regulatory T cell ( β = 0.508, P = .025), CD25 on resting CD4 regulatory T cell ( β = 0.496, P = .018), CD4 on CD39 + activated CD4 regulatory T cell ( β = 0.445, P = .048), CD3 on CD28 + CD45RA − CD8 + T cell ( β = 0.498, P = .024), CD3 on effector memory CD4 + T cell ( β = 0.614, P = .007), and CD3 on activated CD4 regulatory T cell( β = 0.444, P = .049). Using genetic methods, this study uncovers the close relationship between immune cells and PE, which provides guidance for future clinical trials.
- Open Access
- 6 November 2025
- Published by Springer Nature in BMC Medical Informatics and Decision Making
- Vol. 25 (1) , 1-10
Risk stratification and prediction of emergency delivery in early-onset preeclampsia using machine learning
Early-onset preeclampsia (PE) poses significant risks for maternal and fetal outcomes, particularly when emergency delivery is required. This study aimed to develop machine learning models to predict emergency delivery within 48 h of PE diagnosis, facilitating timely clinical interventions. We analyzed a retrospective cohort of 648 singleton pregnancies diagnosed with PE at Fujian Maternal and Child Health Hospital from 2014 to 2024, with gestational ages ranging from 28 to 34 weeks. Patients were stratified into emergency delivery (≤ 48 h post-diagnosis, n = 174) and non-emergency groups (n = 474). Feature selection was performed via univariate analysis, collinearity testing, and logistic regression, yielding 16 predictors. Eight machine learning models (logistic regression, naïve Bayes, XGBoost, LightGBM, SVM, GBDT, MLP, elastic net) were trained and evaluated for discriminative power and calibration. SHAP analysis was employed to interpret model predictions. XGBoost demonstrated superior performance (testing AUROC: 0.908) with excellent calibration, while GBDT achieved high AUROC (0.931) but poorer calibration. SHAP analysis identified CRP, D-dimer, and hypoproteinemia as the most influential predictors. Machine learning models effectively predict emergency delivery in early-onset PE using clinically interpretable features. Integration into obstetric practice may enhance risk stratification, though prospective validation is warranted.
- 6 November 2025
- Published by Springer Nature in BMC Medical Informatics and Decision Making
- Vol. 25 (1) , 1-10
Open AccessRisk stratification and prediction of emergency delivery in early-onset preeclampsia using machine learning
Early-onset preeclampsia (PE) poses significant risks for maternal and fetal outcomes, particularly when emergency delivery is required. This study aimed to develop machine learning models to predict emergency delivery within 48 h of PE diagnosis, facilitating timely clinical interventions. We analyzed a retrospective cohort of 648 singleton pregnancies diagnosed with PE at Fujian Maternal and Child Health Hospital from 2014 to 2024, with gestational ages ranging from 28 to 34 weeks. Patients were stratified into emergency delivery (≤ 48 h post-diagnosis, n = 174) and non-emergency groups (n = 474). Feature selection was performed via univariate analysis, collinearity testing, and logistic regression, yielding 16 predictors. Eight machine learning models (logistic regression, naïve Bayes, XGBoost, LightGBM, SVM, GBDT, MLP, elastic net) were trained and evaluated for discriminative power and calibration. SHAP analysis was employed to interpret model predictions. XGBoost demonstrated superior performance (testing AUROC: 0.908) with excellent calibration, while GBDT achieved high AUROC (0.931) but poorer calibration. SHAP analysis identified CRP, D-dimer, and hypoproteinemia as the most influential predictors. Machine learning models effectively predict emergency delivery in early-onset PE using clinically interpretable features. Integration into obstetric practice may enhance risk stratification, though prospective validation is warranted.
- Open Access
- 1 August 2025
- Published by Taylor & Francis in International Journal of General Medicine
- Vol. ume 18, 4195-4207
A Multi-Algorithm Machine Learning Model for Predicting the Risk of Preterm Birth in Patients with Early-Onset Preeclampsia
Purpose: To analyze the risk factors for preterm birth in patients with early-onset preeclampsia (EOPE) based on multi-algorithm machine learning and to construct a predictive model to explore the predictive value of the model. Methods: A retrospective analysis was conducted on 442 EOPE patients from a single tertiary center, divided into preterm birth (< 37 weeks, n=358) and term-born (≥ 37 weeks, n=84) groups. Univariate analysis, random forest importance assessment, lasso regression combined with multivariate regression analysis were used for feature evaluation. Eight machine learning models were trained (70% data) and validated (30% data). A Stacking ensemble model was constructed, and SHapley Additive exPlanations (SHAP) was used for feature interpretation. Results: The area under the receiver operating characteristic curve (AUROC) for predicting preterm birth in EOPE patients using Logistic Regression, Gaussian Naive Bayes, Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine, Support Vector Machine (SVM), Gradient Boosting Decision Tree (GBDT), Multi-Layer Perceptron, and Elastic Net were 0.763, 0.712, 0.821, 0.832, 0.821, 0.842, 0.784, and 0.763, respectively. The Stacking model (XGBoost+GBDT+SVM) achieved superior performance (AUROC=0.865). Three independent risk factors were identified: fetal growth restriction (aOR=3.50, p = 0.047), serum cystatin C (aOR=11.27, p = 0.018), and C-reactive protein (aOR=1.37, p < 0.001). SHAP analysis revealed GBDT as the top contributor to Stacking predictions, with microalbunminuria (GBDT, XGBoost) and age (SVM) being the most influential features. Conclusion: Machine learning models can serve as reliable assessment tools for predicting the risk of preterm birth in patients with EOPE. The ensemble prediction model demonstrates the best predictive performance, helping obstetricians identify high-risk patients and perform early intervention to improve perinatal outcomes.
- 1 August 2025
- Published by Taylor & Francis in International Journal of General Medicine
- Vol. ume 18, 4195-4207
Open AccessA Multi-Algorithm Machine Learning Model for Predicting the Risk of Preterm Birth in Patients with Early-Onset Preeclampsia
Purpose: To analyze the risk factors for preterm birth in patients with early-onset preeclampsia (EOPE) based on multi-algorithm machine learning and to construct a predictive model to explore the predictive value of the model. Methods: A retrospective analysis was conducted on 442 EOPE patients from a single tertiary center, divided into preterm birth (< 37 weeks, n=358) and term-born (≥ 37 weeks, n=84) groups. Univariate analysis, random forest importance assessment, lasso regression combined with multivariate regression analysis were used for feature evaluation. Eight machine learning models were trained (70% data) and validated (30% data). A Stacking ensemble model was constructed, and SHapley Additive exPlanations (SHAP) was used for feature interpretation. Results: The area under the receiver operating characteristic curve (AUROC) for predicting preterm birth in EOPE patients using Logistic Regression, Gaussian Naive Bayes, Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine, Support Vector Machine (SVM), Gradient Boosting Decision Tree (GBDT), Multi-Layer Perceptron, and Elastic Net were 0.763, 0.712, 0.821, 0.832, 0.821, 0.842, 0.784, and 0.763, respectively. The Stacking model (XGBoost+GBDT+SVM) achieved superior performance (AUROC=0.865). Three independent risk factors were identified: fetal growth restriction (aOR=3.50, p = 0.047), serum cystatin C (aOR=11.27, p = 0.018), and C-reactive protein (aOR=1.37, p < 0.001). SHAP analysis revealed GBDT as the top contributor to Stacking predictions, with microalbunminuria (GBDT, XGBoost) and age (SVM) being the most influential features. Conclusion: Machine learning models can serve as reliable assessment tools for predicting the risk of preterm birth in patients with EOPE. The ensemble prediction model demonstrates the best predictive performance, helping obstetricians identify high-risk patients and perform early intervention to improve perinatal outcomes.
- Open Access
- 24 July 2025
- Published by Springer Nature in BMC Pregnancy and Childbirth
- Vol. 25 (1) , 1-12
Organ-system-based subclassification of preeclampsia using machine learning predicts pregnancy outcomes
Preeclampsia (PE) is a complex disorder with significant variability in organ involvement. It remains unclear whether machine learning can identify organ-system-based subclasses of PE. This study aimed to identify novel subclasses of PE using organ-system biomarkers to improve pregnancy outcome prediction. We retrospectively analyzed clinical data from PE patients at Fujian Maternity and Child Health Hospital. K-means clustering was applied using organ system function indicators, with 10-fold cross-validation. Functional indicators and pregnancy outcomes were compared across subclasses. Heatmap and sankey diagrams were used to reveal the distribution of patients across combined organ system clusters. The analysis included 7,531 PE patients treated between 2013 and 2023. 10-fold cross-validation confirmed clustering robustness with mean ARI of 0.8806 ± 0.0099 and NMI of 0.7800 ± 0.0123. Three heart function clusters were identified using five indicators, with H-Cluster 1 showing the poorest heart function and the highest complication rates. Five kidney clusters were determined from ten indicators. K-Cluster 1 and K-Cluster 5 showed distinct biomarker patterns but similar complication rates (P > 0.05). Liver function analysis using thirteen indicators revealed four clusters. L-Cluster 1 exhibited elevated liver enzymes and bilirubin with higher severe PE and intrahepatic cholestasis rates, whereas L-Cluster 3 had lower protein levels but higher anemia, fetal distress and hemorrhage incidence (P < 0.05). Five coagulation clusters were determined from nine indicators, showing significant differences in indicators and complication rates (P < 0.05). Heatmap and sankey diagram analyses revealed significant overlap between high-risk clusters, with the most frequent combination being H-Cluster 1, K-Cluster 1, L-Cluster 1 and C-Cluster 5. Machine learning identified distinct PE subclasses based on organ system dysfunction patterns, each demonstrating unique pregnancy outcomes. This suggests potential clinical utility of computational approaches for PE subclassification and generates hypotheses for further investigation of its biological mechanisms.
- 24 July 2025
- Published by Springer Nature in BMC Pregnancy and Childbirth
- Vol. 25 (1) , 1-12
Open AccessOrgan-system-based subclassification of preeclampsia using machine learning predicts pregnancy outcomes
Preeclampsia (PE) is a complex disorder with significant variability in organ involvement. It remains unclear whether machine learning can identify organ-system-based subclasses of PE. This study aimed to identify novel subclasses of PE using organ-system biomarkers to improve pregnancy outcome prediction. We retrospectively analyzed clinical data from PE patients at Fujian Maternity and Child Health Hospital. K-means clustering was applied using organ system function indicators, with 10-fold cross-validation. Functional indicators and pregnancy outcomes were compared across subclasses. Heatmap and sankey diagrams were used to reveal the distribution of patients across combined organ system clusters. The analysis included 7,531 PE patients treated between 2013 and 2023. 10-fold cross-validation confirmed clustering robustness with mean ARI of 0.8806 ± 0.0099 and NMI of 0.7800 ± 0.0123. Three heart function clusters were identified using five indicators, with H-Cluster 1 showing the poorest heart function and the highest complication rates. Five kidney clusters were determined from ten indicators. K-Cluster 1 and K-Cluster 5 showed distinct biomarker patterns but similar complication rates (P > 0.05). Liver function analysis using thirteen indicators revealed four clusters. L-Cluster 1 exhibited elevated liver enzymes and bilirubin with higher severe PE and intrahepatic cholestasis rates, whereas L-Cluster 3 had lower protein levels but higher anemia, fetal distress and hemorrhage incidence (P < 0.05). Five coagulation clusters were determined from nine indicators, showing significant differences in indicators and complication rates (P < 0.05). Heatmap and sankey diagram analyses revealed significant overlap between high-risk clusters, with the most frequent combination being H-Cluster 1, K-Cluster 1, L-Cluster 1 and C-Cluster 5. Machine learning identified distinct PE subclasses based on organ system dysfunction patterns, each demonstrating unique pregnancy outcomes. This suggests potential clinical utility of computational approaches for PE subclassification and generates hypotheses for further investigation of its biological mechanisms.
- Open Access
- 12 March 2025
- Published by Frontiers Media SA in Frontiers in Genetics
- Vol. 16, 1500061
Identification and validation of key biomarkers associated with immune and oxidative stress for preeclampsia by WGCNA and machine learning
Background: Preeclampsia (PE), a major obstetric disorder marked by dysfunction in both placental and maternal vascular systems, continues to pose critical challenges in global maternal healthcare. This multisystem pregnancy complication contributes significantly to adverse perinatal outcomes and remains a leading cause of pregnancy-related morbidity worldwide. However, the available treatment options at present remain restricted. Our investigation employs an integrative bioinformatics approach to elucidate critical molecular signatures linked to the interplay between immunological dysregulation and oxidative stress mechanisms in PE pathogenesis.Methods: In this study, we sourced the dataset from the GEO database with the aim of pinpointing differentially expressed genes (DEGs) between PE samples and control samples. Genes associated with oxidative stress were procured from the Genecards database. Next, we employed a comprehensive approach. This involved integrating WGCNA, GO and KEGG pathway analyses, constructing PPI networks, applying machine learning algorithms, performing gene GSEA, and conducting immune infiltration analysis to identify the key hub genes related to oxidative stress. Diagnostic potential of candidate biomarkers was quantitatively assessed through ROC curve modeling. Additionally, we constructed a miRNA - gene regulatory network for the identified diagnostic genes and predicted potential candidate drugs. In the final step, we validated the significant hub gene using independent external datasets, the hypoxia model of the HTR-8/SVneo cell line, and human placental tissue samples.Results: At last, leptin (LEP) was identified as a core gene through screening and was found to be upregulated. The results of quantitative real-time polymerase chain reaction (qRT -PCR) and immunohistochemistry validation were consistent with those obtained from the datasets. KEGG analysis revealed that LEP was significantly enriched in “allograft rejection,” “antigen processing,” “ECM receptor interaction” and “graft versus host disease.” GO analysis revealed that LEP was involved in biological processes such as “antigen processing and presentation,” “peptide antigen assembly with MHC protein complex,” “complex of collagen trimers,” “MHC class II protein complex” and “mitochondrial protein containing complex.” Moreover, immune cell analysis indicated that T follicular helper cells, plasmacytoid dendritic cells, neutrophils, and activated dendritic cells were positively correlated with LEP expression, whereas γδT cells, eosinophils, and central memory CD4+ T cells showed a negative correlation. These findings suggest that LEP influences the immune microenvironment of PE through its interaction with arious immune cells. In addition, 28 miRNAs and 15 drugs were predicted to target LEP. Finally, the overexpression of LEP was verified using independent external datasets, the hypoxia model of the HTR-8/SVneo cell line, and human placental tissue.Conclusion: Through an integrated analytical framework employing WGCNA coupled with three distinct machine learning-driven phenotypic classification models, we discovered a pivotal regulatory gene. This gene has the potential to act as a novel diagnostic biomarker for PE. Moreover, it can be considered as a promising target for drug development related to PE. Notably, it shows a strong correlation with the immune microenvironment, suggesting its crucial role in the complex pathophysiological processes underlying PE.
- 12 March 2025
- Published by Frontiers Media SA in Frontiers in Genetics
- Vol. 16, 1500061
Open AccessIdentification and validation of key biomarkers associated with immune and oxidative stress for preeclampsia by WGCNA and machine learning
Background: Preeclampsia (PE), a major obstetric disorder marked by dysfunction in both placental and maternal vascular systems, continues to pose critical challenges in global maternal healthcare. This multisystem pregnancy complication contributes significantly to adverse perinatal outcomes and remains a leading cause of pregnancy-related morbidity worldwide. However, the available treatment options at present remain restricted. Our investigation employs an integrative bioinformatics approach to elucidate critical molecular signatures linked to the interplay between immunological dysregulation and oxidative stress mechanisms in PE pathogenesis.Methods: In this study, we sourced the dataset from the GEO database with the aim of pinpointing differentially expressed genes (DEGs) between PE samples and control samples. Genes associated with oxidative stress were procured from the Genecards database. Next, we employed a comprehensive approach. This involved integrating WGCNA, GO and KEGG pathway analyses, constructing PPI networks, applying machine learning algorithms, performing gene GSEA, and conducting immune infiltration analysis to identify the key hub genes related to oxidative stress. Diagnostic potential of candidate biomarkers was quantitatively assessed through ROC curve modeling. Additionally, we constructed a miRNA - gene regulatory network for the identified diagnostic genes and predicted potential candidate drugs. In the final step, we validated the significant hub gene using independent external datasets, the hypoxia model of the HTR-8/SVneo cell line, and human placental tissue samples.Results: At last, leptin (LEP) was identified as a core gene through screening and was found to be upregulated. The results of quantitative real-time polymerase chain reaction (qRT -PCR) and immunohistochemistry validation were consistent with those obtained from the datasets. KEGG analysis revealed that LEP was significantly enriched in “allograft rejection,” “antigen processing,” “ECM receptor interaction” and “graft versus host disease.” GO analysis revealed that LEP was involved in biological processes such as “antigen processing and presentation,” “peptide antigen assembly with MHC protein complex,” “complex of collagen trimers,” “MHC class II protein complex” and “mitochondrial protein containing complex.” Moreover, immune cell analysis indicated that T follicular helper cells, plasmacytoid dendritic cells, neutrophils, and activated dendritic cells were positively correlated with LEP expression, whereas γδT cells, eosinophils, and central memory CD4+ T cells showed a negative correlation. These findings suggest that LEP influences the immune microenvironment of PE through its interaction with arious immune cells. In addition, 28 miRNAs and 15 drugs were predicted to target LEP. Finally, the overexpression of LEP was verified using independent external datasets, the hypoxia model of the HTR-8/SVneo cell line, and human placental tissue.Conclusion: Through an integrated analytical framework employing WGCNA coupled with three distinct machine learning-driven phenotypic classification models, we discovered a pivotal regulatory gene. This gene has the potential to act as a novel diagnostic biomarker for PE. Moreover, it can be considered as a promising target for drug development related to PE. Notably, it shows a strong correlation with the immune microenvironment, suggesting its crucial role in the complex pathophysiological processes underlying PE.
- Open Access
- 24 January 2025
- Published by Wiley in The Journal of Clinical Hypertension
- Vol. 27 (1) , e14982
Identification and Immunological Characterization of Cuproptosis Related Genes in Preeclampsia Using Bioinformatics Analysis and Machine Learning
Preeclampsia (PE) is a pregnancy‐specific disorder characterized by an unclearly understood pathogenesis and poses a great threat to maternal and fetal safety. Cuproptosis, a novel form of cellular death, has been implicated in the advancement of various diseases. However, the role of cuproptosis and immune‐related genes in PE is unclear. The current study aims to elucidate the gene expression matrix and immune infiltration patterns of cuproptosis‐related genes (CRGs) in the context of PE. The GSE98224 dataset was obtained from the Gene Expression Omnibus (GEO) database and utilized as the internal training set. Based on the GSE98224 dataset, we explored the differentially expressed cuproptosis related genes (DECRGs) and immunological composition. We identified 10 DECRGs conducted Gene Ontology (GO) function, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses, and a protein–protein interaction (PPI) network. Furthermore, patients with PE were categorized into two distinct clusters, and an investigation was conducted to examine the status of immune cell infiltration. Additionally, the application of Weighted Gene Co‐expression Network Analysis (WGCNA) was utilized to differentiate modules consisting of co‐expressed genes and conduct clustering analysis. The intersecting genes were obtained by intersecting differently expressed genes in PE and PE clusters. The most precise forecasting model was chosen by evaluating the effectiveness of four machine learning models. The ResNet model was established to score the hub genes. The prediction accuracy was assessed by receiver operating characteristic (ROC) curves and an external dataset. We successfully identified five key DECREGs and two pathological clusters in PE, each with distinct immune profiles and biological characteristics. Subsequently, the RF model was deemed the most optimal model for the identification of PE with a large area under the curve (AUC = 0.733). The five genes that ranked highest in the RF machine learning model were considered to be predictor genes. The calibration curve demonstrated a high level of accuracy in aligning the predicted outcomes with the actual outcomes. We validate the ResNet model using the ROC curve with the area under the curve (AUC = 0.82). Cuproptosis and immune infiltration may play an important role in the pathogenesis of PE. The present study elucidated that GSTA4, KCNK5, APLNR, IKZF2, and CAP2 may be potential markers of cuproptosis‐associated PE and are considered to play a significant role in the initiation and development of cuproptosis‐induced PE.
- 24 January 2025
- Published by Wiley in The Journal of Clinical Hypertension
- Vol. 27 (1) , e14982
Open AccessIdentification and Immunological Characterization of Cuproptosis Related Genes in Preeclampsia Using Bioinformatics Analysis and Machine Learning
Preeclampsia (PE) is a pregnancy‐specific disorder characterized by an unclearly understood pathogenesis and poses a great threat to maternal and fetal safety. Cuproptosis, a novel form of cellular death, has been implicated in the advancement of various diseases. However, the role of cuproptosis and immune‐related genes in PE is unclear. The current study aims to elucidate the gene expression matrix and immune infiltration patterns of cuproptosis‐related genes (CRGs) in the context of PE. The GSE98224 dataset was obtained from the Gene Expression Omnibus (GEO) database and utilized as the internal training set. Based on the GSE98224 dataset, we explored the differentially expressed cuproptosis related genes (DECRGs) and immunological composition. We identified 10 DECRGs conducted Gene Ontology (GO) function, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses, and a protein–protein interaction (PPI) network. Furthermore, patients with PE were categorized into two distinct clusters, and an investigation was conducted to examine the status of immune cell infiltration. Additionally, the application of Weighted Gene Co‐expression Network Analysis (WGCNA) was utilized to differentiate modules consisting of co‐expressed genes and conduct clustering analysis. The intersecting genes were obtained by intersecting differently expressed genes in PE and PE clusters. The most precise forecasting model was chosen by evaluating the effectiveness of four machine learning models. The ResNet model was established to score the hub genes. The prediction accuracy was assessed by receiver operating characteristic (ROC) curves and an external dataset. We successfully identified five key DECREGs and two pathological clusters in PE, each with distinct immune profiles and biological characteristics. Subsequently, the RF model was deemed the most optimal model for the identification of PE with a large area under the curve (AUC = 0.733). The five genes that ranked highest in the RF machine learning model were considered to be predictor genes. The calibration curve demonstrated a high level of accuracy in aligning the predicted outcomes with the actual outcomes. We validate the ResNet model using the ROC curve with the area under the curve (AUC = 0.82). Cuproptosis and immune infiltration may play an important role in the pathogenesis of PE. The present study elucidated that GSTA4, KCNK5, APLNR, IKZF2, and CAP2 may be potential markers of cuproptosis‐associated PE and are considered to play a significant role in the initiation and development of cuproptosis‐induced PE.
- Open Access
- 1 September 2023
- Published by Wolters Kluwer Health in Medicine
- Vol. 102 (35) , e34551
The application of rapid response team in category 1 emergency caesarean section teaching for OBGYN residents in the delivery room
Category 1 cesarean section (CS) can be a life-saving procedure when there is immediate threat to the life of the woman or fetus. However, category 1 CS is a challenge for obstetrics and gynecology residents, and it is necessary to establish an effective and straightforward teaching strategy. This study aimed to evaluate the efficiency of rapid response team (RRT) on category 1 CS teaching for obstetrics and gynecology residents in the delivery room. A total of 142 residents who underwent standardized residency training programs in the delivery room were divided into a RRT teaching group and a traditional response (TR) teaching group. In the RRT teaching group, Category 1 emergency CS teaching was started and explored by rapid response team. The training included both theoretical and practical components. After the training, decision-to-delivery interval (DDI), neonatal Apgar score, operation time and rate of postpartum hemorrhage were compared. A questionnaire on the subjective assessment of various aspects of the program was conducted at the end of the training period. The DDI in minutes in the RRT teaching group (n = 72) was significantly shorter than that of the TR teaching group (n = 70) (11.83 ± 4.16 vs 13.56 ± 5.47, P = .0364). The score of satisfaction from residents in the RRT teaching group was significantly higher than that of the TR group [7 (6, 9) vs 9 (7, 10), P = .0154]. Compared with the TR teaching group, more residents thought their clinical skills have been improved (94.29% vs 100%, P = .0396) and willing to recommend their training method to others (91.43% vs 100%, P = .0399) in the RRT teaching group. However, no significant differences were observed in the incidence of postpartum hemorrhage between the 2 groups. RRT teaching is beneficial in the standardized training and teaching of residents in the delivery room. It improves the DDI of category 1 emergency cesarean section and the degree of satisfaction.
- 1 September 2023
- Published by Wolters Kluwer Health in Medicine
- Vol. 102 (35) , e34551
Open AccessThe application of rapid response team in category 1 emergency caesarean section teaching for OBGYN residents in the delivery room
Category 1 cesarean section (CS) can be a life-saving procedure when there is immediate threat to the life of the woman or fetus. However, category 1 CS is a challenge for obstetrics and gynecology residents, and it is necessary to establish an effective and straightforward teaching strategy. This study aimed to evaluate the efficiency of rapid response team (RRT) on category 1 CS teaching for obstetrics and gynecology residents in the delivery room. A total of 142 residents who underwent standardized residency training programs in the delivery room were divided into a RRT teaching group and a traditional response (TR) teaching group. In the RRT teaching group, Category 1 emergency CS teaching was started and explored by rapid response team. The training included both theoretical and practical components. After the training, decision-to-delivery interval (DDI), neonatal Apgar score, operation time and rate of postpartum hemorrhage were compared. A questionnaire on the subjective assessment of various aspects of the program was conducted at the end of the training period. The DDI in minutes in the RRT teaching group (n = 72) was significantly shorter than that of the TR teaching group (n = 70) (11.83 ± 4.16 vs 13.56 ± 5.47, P = .0364). The score of satisfaction from residents in the RRT teaching group was significantly higher than that of the TR group [7 (6, 9) vs 9 (7, 10), P = .0154]. Compared with the TR teaching group, more residents thought their clinical skills have been improved (94.29% vs 100%, P = .0396) and willing to recommend their training method to others (91.43% vs 100%, P = .0399) in the RRT teaching group. However, no significant differences were observed in the incidence of postpartum hemorrhage between the 2 groups. RRT teaching is beneficial in the standardized training and teaching of residents in the delivery room. It improves the DDI of category 1 emergency cesarean section and the degree of satisfaction.