PPARG as a Sepsis Diagnostic Biomarker: Integrated Transcriptomics and Validation

PPARG was identified as a diagnostic biomarker for sepsis with an AUC of 0.994. This dataset contains whole-blood transcriptomic data from two cohorts, GSE236713 and GSE65682, used to discover and validate biomarkers linked to CD14/NF-κB signaling. Yingyi Ji published the data on figshare in 2026.

Use Cases

• Train a sepsis diagnostic classifier based on the five identified feature genes (MMP9, PPARG, C1QC, MS4A4A, ARG1).
• Validate biomarker associations with clinical outcomes based on reported correlations with SOFA scores and ICU risk.
• Analyze immune cell composition in sepsis samples using the CIBERSORT method mentioned in the description.
• Investigate pathway activity in sepsis using the Gene Set Enrichment Analysis (GSEA) results for NF-κB-related pathways.

Strengths

• External validation was performed using an independent cohort (GSE65682).
• Multiple machine learning algorithms were compared, with an artificial neural network achieving an AUC of 0.974 internally and 0.953 externally.
• Experimental validation was conducted using LPS-stimulated macrophages with CD14 knockout/overexpression.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect bias inherent to the specific transcriptomic cohorts used.

Provenance

Source

Gene Expression Omnibus (GEO) datasets GSE236713 and GSE65682.

Collection Method

Whole-blood transcriptomic data analyzed with differential expression, WGCNA, protein-protein interaction, LASSO, and random forest.

Freshness

Last updated 2026-05-28 04:43:37; freshness should be verified.