Machine Learning-Integrated Network Toxicology Uncovers Glioma Targets of DEHP

77 overlapping genes were identified from an integrative analysis of glioma and potential DEHP targets. The study by Ren Li, published in 2026, prioritized 12 key genes using 127 machine learning models and validated findings in U87 cells. The data supports a 12-gene diagnostic model with an external validation AUC of 0.994.

Use Cases

• Prioritizing molecular targets for environmental toxins based on integrated network toxicology and machine learning methods described.
• Building diagnostic models for glioma based on the 12-gene signature identified in the study.
• Assessing prognostic risk scores for glioma patients based on DEHP-related gene expression patterns.
• Validating molecular docking and dynamics simulations for toxin-protein interactions as performed in the research.

Strengths

• Integrates data from four established databases (CHEMBL, CTD, SwissTargetPrediction, PharmMapper) and public repositories (GEO, TCGA).
• Employs a substantial machine learning pipeline of 127 models to prioritize core genes.
• External validation was performed across multiple cohorts (CGGA_325, CGGA_693, GSE16011) with a reported pooled AUC of 0.994.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The 1.1 MB file size suggests a small dataset, likely containing summary results rather than raw data.

Provenance

Source

figshare

Collection Method

Potential DEHP targets were identified through databases; glioma-related genes were screened using differential expression analysis and WGCNA on GEO and TCGA datasets.

Freshness

Last updated 2026-05-07 22:02:33