Mouse Transcriptomic Biomarkers for Radiation Dose Reconstruction via Machine Learning

172 robust biomarkers were identified from 25,654 genes in a mouse model for radiation biodosimetry. A stacked ensemble machine learning model achieved dose reconstruction across 0–12 Gy with an R² of 0.952 and RMSE of 0.938 Gy. The dataset, created by Yucheng Wang and last updated in 2026, provides a preclinical foundation for human radiation exposure assessment.

Use Cases

• Training ML models for radiation dose reconstruction based on gene expression biomarkers.
• Screening for robust radiation-responsive biomarkers based on transcriptomic sequencing data.
• Validating biomarker panels for field-deployable diagnostic platforms based on qRT-PCR results.

Strengths

• High-accuracy dose reconstruction model reported with an R² of 0.952 and RMSE of 0.938 Gy.
• Biomarker screening process reduced an initial pool of 25,654 genes to a refined panel of 15 key markers.
• Methodology validated with qRT-PCR, confirming reliability for translational research.

Limitations

• Data is derived from a female C57BL/6J mouse model; direct applicability to human subjects requires further translational research.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

figshare

Collection Method

Transcriptomic sequencing of blood samples from irradiated mice, followed by differential expression analysis and machine learning.

Freshness

Last updated 2026-05-18 14:02:52; freshness should be verified.