Pediatric Sepsis Biomarker Candidates from Multi-Model Machine Learning Analysis

A 2026 study by Weidong Ye integrated six bulk RNA-seq datasets containing 497 pediatric sepsis patients and 116 healthy controls. The research identified 237 high-confidence biomarker genes, with RORA and GPR183 highlighted across 14 machine learning models. The dataset supports the development of diagnostic tools and targeted therapies for this condition.

Use Cases

• Validate the 237 identified biomarker genes, such as RORA and GPR183, in independent sepsis cohorts.
• Analyze correlations between biomarker expression levels and immune cell infiltration scores for B cells, CD8+ T cells, neutrophils, and monocytes.
• Build classification models to distinguish sepsis patients from healthy controls using the gene expression signatures from the integrated RNA-seq data.
• Investigate the functional roles of top biomarkers in biological processes like transcription, translation, and cellular senescence.
• Compare biomarker performance across the 14 different machine learning models used in the original screening process.

Strengths

• Integrates data from 613 total samples (497 patients, 116 controls) across six source datasets.
• Identifies 237 high-confidence biomarker candidates validated through a multi-model machine learning approach.
• Includes functional and immune infiltration analysis linking biomarkers to specific biological processes and cell types.

Limitations

• The dataset is small in file size (157.4 KB), indicating it likely contains summary results rather than raw expression data.
• Sample provenance and specific patient demographics from the six integrated studies are not detailed.
• The analysis is based on bulk RNA-seq, which lacks single-cell resolution for the observed immune cell changes.

Provenance

Source

figshare, authored by Weidong Ye.

Collection Method

Integration and re-analysis of six existing bulk RNA-seq datasets related to pediatric sepsis.

Time Range

null

Freshness

Last updated in April 2026.

Geography

null