Related Experiment Video
Updated: Aug 6, 2026

Supervised Machine Learning for Semi-Quantification of Extracellular DNA in Glomerulonephritis
Published on: June 18, 2020
WGCNA combined with machine learning identifies histone deacetylation-related diagnostic features for pediatric
Ji Yao1, Zhongbin Lai1, Di Yu1
1Department of Anesthesiology, Medical and Health Group, Cixi Third People's Hospital, Ningbo 315324, China.
Abstract:
Pediatric septic shock is a severe form of sepsis with high mortality. Histone deacetylation is involved in sepsis-related disorders. This study investigated the diagnostic value of histone deacetylation-related genes in pediatric septic shock. Three Gene Expression Omnibus datasets (GSE26378, GSE26440, and GSE13904) were analyzed. Differentially expressed genes from GSE26378 and GSE26440 were intersected with genes in key modules identified by weighted gene co-expression network analysis. Least absolute shrinkage and selection operator regression and random forest were used to identify hub genes, and receiver operating characteristic curves assessed diagnostic performance. Quantitative real-time polymerase chain reaction validated candidate gene expression. Immunoprecipitation preliminarily evaluated protein acetylation changes after trichostatin A treatment. Immune infiltration was analyzed, and related transcription factors and miRNAs were predicted. C9orf84, ASAP1-IT1, and CREB5 were identified as hub genes. All were upregulated in pediatric septic shock and showed favorable diagnostic performance in the training and validation datasets. Trichostatin A increased the acetylation levels and reduced the protein levels of C9orf84 and CREB5. Neutrophils, macrophages, and regulatory T cells were more abundant in pediatric septic shock than in controls. Multiple potentially interacting transcription factors and miRNAs were also identified. C9orf84, ASAP1-IT1, and CREB5 may serve as diagnostic biomarkers and therapeutic targets for pediatric septic shock.
