Saved in:
| Main Author: | |
|---|---|
| Format: | Recurso digital |
| Language: | |
| Published: |
Zenodo
2026
|
| Online Access: | https://doi.org/10.5281/zenodo.20262520 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866902026096476160 |
|---|---|
| author | Roozbeh Heidarzadehpilehrood |
| author_facet | Roozbeh Heidarzadehpilehrood |
| contents | <h2>Overview</h2> <p>This is the <strong>initial public release (v1.1)</strong> of the analysis code accompanying the manuscript on machine learning–based prognostic biomarker discovery in ** High-Grade Serous Carcinoma (HGSC)**.</p> <p>The repository contains a fully scripted R pipeline for:</p> <ul> <li>GEO dataset acquisition and preprocessing</li> <li>Differential expression (DEG) analysis using <code>limma</code></li> <li>Cross-cohort consensus DEG identification</li> <li>Machine learning model training under strict <strong>Leave-One-Dataset-Out (LODO)</strong> validation</li> <li>Multi-model feature importance ranking</li> <li>Candidate biomarker selection</li> </ul> <h2>Included scripts</h2> <p>All analysis scripts are located in the <code>scripts/</code> directory:</p> <ul> <li><code>00_setup.R</code> — install and load dependencies</li> <li><code>01_params.R</code> — global parameters and paths</li> <li><code>02_download_geo.R</code> — GEO data download and preprocessing</li> <li><code>03_make_sample_sheets.R</code> — sample sheet construction</li> <li><code>04_deg_mrna_limma.R</code> — DEG analysis (limma)</li> <li><code>05_meta_deg_consensus.R</code> — cross-cohort meta-analytic DEG consensus</li> <li><code>06_ml_diagnostic_LODO.R</code> — ML model training + LODO validation</li> <li><code>07_benchmark_models_STRICT_LODO.R</code> — strict multi-model benchmarking</li> <li><code>08_multimodel_gene_ranking_STRICT_LODO_top500.R</code> — multi-model gene ranking (top 500)</li> <li><code>09_select_candidate_biomarkers_STRICT_LODO.R</code> — final candidate biomarker selection</li> </ul> <h2>Notes</h2> <ul> <li><strong>Data files</strong> (raw GEO data, intermediate <code>.rds</code> / <code>.csv</code>, and large result objects) are <strong>not</strong> included in this release to keep the repository lightweight.</li> <li>All required paths and parameters can be configured in <code>scripts/01_params.R</code>.</li> <li>This release is intended for <strong>reproducibility and peer review</strong>. Further updates (e.g., additional visualizations and figure generation scripts) may be added in future versions.</li> </ul> <h2>Citation</h2> <p>If you use this code, please cite:</p> <p>Heidarzadehpilehrood R. et al. (2026).<br> Machine learning–based discovery of prognostic biomarkers in serous ovarian cancer.<br> GitHub release: <code>v1.1</code>.</p> <p>(Repository DOI to be provided via Zenodo.)</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_20262520 |
| institution | Zenodo |
| language | |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | heidarzadehroozbeh-cmyk/ML-Prognostic_Biomarkers-for-Serous-OvarianCancer: Version 1.1 – Machine Learning Pipeline for HGSC Prognostic Biomarker Discovery Roozbeh Heidarzadehpilehrood <h2>Overview</h2> <p>This is the <strong>initial public release (v1.1)</strong> of the analysis code accompanying the manuscript on machine learning–based prognostic biomarker discovery in ** High-Grade Serous Carcinoma (HGSC)**.</p> <p>The repository contains a fully scripted R pipeline for:</p> <ul> <li>GEO dataset acquisition and preprocessing</li> <li>Differential expression (DEG) analysis using <code>limma</code></li> <li>Cross-cohort consensus DEG identification</li> <li>Machine learning model training under strict <strong>Leave-One-Dataset-Out (LODO)</strong> validation</li> <li>Multi-model feature importance ranking</li> <li>Candidate biomarker selection</li> </ul> <h2>Included scripts</h2> <p>All analysis scripts are located in the <code>scripts/</code> directory:</p> <ul> <li><code>00_setup.R</code> — install and load dependencies</li> <li><code>01_params.R</code> — global parameters and paths</li> <li><code>02_download_geo.R</code> — GEO data download and preprocessing</li> <li><code>03_make_sample_sheets.R</code> — sample sheet construction</li> <li><code>04_deg_mrna_limma.R</code> — DEG analysis (limma)</li> <li><code>05_meta_deg_consensus.R</code> — cross-cohort meta-analytic DEG consensus</li> <li><code>06_ml_diagnostic_LODO.R</code> — ML model training + LODO validation</li> <li><code>07_benchmark_models_STRICT_LODO.R</code> — strict multi-model benchmarking</li> <li><code>08_multimodel_gene_ranking_STRICT_LODO_top500.R</code> — multi-model gene ranking (top 500)</li> <li><code>09_select_candidate_biomarkers_STRICT_LODO.R</code> — final candidate biomarker selection</li> </ul> <h2>Notes</h2> <ul> <li><strong>Data files</strong> (raw GEO data, intermediate <code>.rds</code> / <code>.csv</code>, and large result objects) are <strong>not</strong> included in this release to keep the repository lightweight.</li> <li>All required paths and parameters can be configured in <code>scripts/01_params.R</code>.</li> <li>This release is intended for <strong>reproducibility and peer review</strong>. Further updates (e.g., additional visualizations and figure generation scripts) may be added in future versions.</li> </ul> <h2>Citation</h2> <p>If you use this code, please cite:</p> <p>Heidarzadehpilehrood R. et al. (2026).<br> Machine learning–based discovery of prognostic biomarkers in serous ovarian cancer.<br> GitHub release: <code>v1.1</code>.</p> <p>(Repository DOI to be provided via Zenodo.)</p> |
| title | heidarzadehroozbeh-cmyk/ML-Prognostic_Biomarkers-for-Serous-OvarianCancer: Version 1.1 – Machine Learning Pipeline for HGSC Prognostic Biomarker Discovery |
| url | https://doi.org/10.5281/zenodo.20262520 |