Characteristics of Study Cohorts: Pancreatic Cancer miRNA Diagnostic Model Data

801 samples from public datasets were used to develop and validate a 20-microRNA signature diagnostic model for pancreatic cancer. The model was trained on 216 samples and validated on 585 independent samples, achieving cross-validation AUC of 0.87. The dataset, authored by Yanfei Zhu and last updated in May 2026, is shared under a CC-BY-4.0 license.

Use Cases

• Validating miRNA-based diagnostic models for pancreatic cancer based on the described 20-miRNA signature.
• Benchmarking machine learning classifier performance on multi-cohort biomedical data based on the described Random Forest and SHAP analysis.
• Conducting functional enrichment analysis on miRNA targets based on the described cancer-related pathways.
• Assessing model generalizability across RNA-seq and qRT-PCR platforms based on the described external validation strategy.

Strengths

• Includes 801 total samples from multiple public sources, providing a basis for cross-validation.
• Model performance is quantified with specific metrics, including an AUC of 0.87 and sensitivity of 84.7%.
• Dataset is small (5.5 KB) and in XLS format, facilitating quick inspection and download.

Limitations

• Row count and column-level documentation are unknown, which may limit suitability assessment.
• Data is based on retrospective studies, which may limit immediate clinical applicability as noted in the description.

Provenance

Source

Publicly available datasets (GSE59856, GSE85589, TCGA-PAAD, GTEx, GSE128508) aggregated for research.

Collection Method

Samples were aggregated from public repositories for model training and external validation.

Time Range

null

Freshness

Last updated 2026-05-04 17:37:44; freshness should be verified.

Geography

null