Development of immune-derived molecular markers for preeclampsia based on multiple machine learning algorithms.

Area of Science:

• Immunology
• Genomics
• Bioinformatics

Background:

• Preeclampsia (PE) is a critical pregnancy complication with significant maternal and neonatal risks.
• The role of immune dysregulation in PE pathogenesis remains incompletely understood.
• Identifying molecular markers associated with immune infiltration is crucial for therapeutic development.

Purpose of the Study:

• To identify immune-related differentially expressed genes (DEGs) in preeclampsia.
• To develop a predictive model for PE using machine learning and immune biomarkers.
• To explore the relationship between immune cells and molecular signatures in PE.

Main Methods:

• Bioinformatic analysis of gene expression data from the GEO database.
• Differential expression analysis using DESeq2 and limma.
• Machine learning algorithms (LASSO, bagged trees, RF) for biomarker selection and model construction.

Main Results:

• Identified 34 immune source-related DEGs.
• Developed a robust diagnostic forecasting model for PE using ML-derived immune biomarkers.
• Found significant associations between six immune cell types and PE biomarkers.
• Discovered immune-derived hub genes with potential drug-binding capabilities (e.g., alitretinoin).

Conclusions:

• The developed prediction model integrates multiple immune-related biomarkers for enhanced PE prediction.
• Identified biomarkers show potential to outperform existing molecular signatures for PE.
• The study provides insights into immune dysregulation mechanisms in PE, paving the way for clinical applications.