Related Experiment Video
Updated: Jun 14, 2025
07:42
A Data-Driven Approach to Quantifying Immune States in Sepsis
Published on: February 7, 2025
150
Risk assessment with gene expression markers in sepsis development.
Albert Garcia Lopez1, Sascha Schäuble1, Tongta Sae-Ong1
1Department of Microbiome Dynamics, Leibniz Institute for Natural Product Research and Infection Biology (Leibniz-HKI), 07745 Jena, Germany.
Cell Reports. Medicine
|September 4, 2024
Summary
Preoperative gene expression can predict sepsis risk in surgical patients. Machine learning models identify individuals predisposed to developing sepsis, aiding early intervention and patient management.
Area of Science:
- Genomics
- Immunology
- Computational Biology
Background:
- Sepsis is a life-threatening response to infection.
- Predicting sepsis risk in surgical patients remains challenging.
- Individual susceptibility to sepsis varies significantly.
Purpose of the Study:
- To investigate preoperative transcriptomic signatures for predicting postoperative sepsis.
- To identify individual phenotypic predispositions to sepsis development.
- To develop a risk prediction tool for postoperative sepsis.
Main Methods:
- Whole-blood RNA sequencing on preoperative samples from 267 patients.
- Machine learning classification models built on transcriptomic data.
- Validation using quantitative reverse-transcription PCR (RT-qPCR).
Main Results:
- Machine learning models predicted postoperative outcomes, including sepsis, with high accuracy (AUC up to 0.910).
- Models achieved sensitivity and specificity up to 0.767 and 0.804, respectively.
- Identified preoperative transcriptomic signatures associated with sepsis development.
Conclusions:
- Preoperative transcriptomic signatures can predict the risk of developing postoperative sepsis.
- Machine learning models offer a potential tool for sepsis risk prediction in surgical patients.
- Findings suggest an individual predisposition to sepsis that requires further investigation.
