Related Experiment Video
Updated: Dec 10, 2025
07:42
A Data-Driven Approach to Quantifying Immune States in Sepsis
Published on: February 7, 2025
409
Machine learning based refined differential gene expression analysis of pediatric sepsis
Mostafa Abbas1, Yasser El-Manzalawy2,3
1Department of Imaging Science and Innovation, Geisinger Health System, Danville, PA, 17822, USA.
BMC Medical Genomics
|August 30, 2020
Summary
This study introduces a novel machine learning method to identify key gene biomarkers from differential expression analysis. A 10-gene signature effectively predicts pediatric sepsis mortality with high accuracy.
Area of Science:
- Transcriptomics
- Bioinformatics
- Machine Learning
Background:
- Differential expression (DE) analysis identifies gene expression changes in biological conditions.
- Differentially expressed genes (DEGs) are crucial for functional pathway analysis and biomarker discovery.
- Current methods for identifying DEGs as biomarkers can be enhanced.
Purpose of the Study:
- To develop a novel approach for identifying robust biomarkers from DEGs.
- To refine DE analysis using machine learning for improved biomarker prioritization.
- To discover a gene signature for predicting pediatric sepsis mortality.
Main Methods:
- Utilized Minimum Redundancy Maximum Relevance (MRMR) criteria for re-ranking DEGs.
- Employed a repeated cross-validation feature selection procedure.
- Applied the method to gene expression profiles of pediatric sepsis patients.
Main Results:
- Identified 108 DEGs from 199 children with sepsis and septic shock.
- Developed a 10-gene signature for predicting pediatric sepsis mortality.
- Achieved an Area Under the ROC Curve (AUC) score of 0.89 for the signature's predictive performance.
Conclusions:
- Machine learning refinement of DE analysis is a powerful tool for biomarker discovery.
- The identified 10-gene signature shows promise for reliable diagnosis and prognosis of sepsis.
- This approach facilitates the development of novel sepsis biomarkers.
