SCAPeSCLC is a harmonized multi-level transcriptomic and clinical resource derived from the publicly available GEO datasets GSE261345 and GSE261348, originating from the CANTABRICO and IMfirst cohorts of patients with extensive-stage small cell lung cancer (ES-SCLC).
This repository contains the R scripts and supporting datasets used to generate Bayesian pathway posterior estimates, perform gene- and pathway-level survival analyses, assess Cox proportional hazards model assumptions, and generate comprehensive diagnostic atlases for both gene expression and biological pathway activity.
The repository accompanies SCAPeSCLC v1.3.5, the published dataset and associated data paper.
Major additions include:
- Gene-level diagnostic atlases (unadjusted and confounder-adjusted)
- Biological pathway diagnostic atlases (unadjusted and confounder-adjusted)
- Proportional hazards assumption testing for all gene and pathway Cox models
- Bayesian estimation of patient-level Cancer Transcriptome Atlas pathway activities
- Complete SCAPeSCLC dataset provided in CSV and consolidated XLSX formats
- Zenodo dataset (archival repository): https://doi.org/10.5281/zenodo.19897644
- Data paper:
Shirvaliloo M. SCAPeSCLC: An Integrated Spatial Transcriptomic and Bayesian Pathway Enrichment Dataset for Survival Modeling in Extensive-Stage Small Cell Lung Cancer. Data. 2026;11(7):152. https://doi.org/10.3390/data11070152
SCAPeSCLC
├── data/
│ ├── D5_scaled_gene_expression.csv
│ ├── D10_ROI_CTA_Zscores.csv
│ └── D13_patient_BP_posteriors.csv
│
├── dataset/
│ ├── csv/
│ │ ├── D01_patient_demographics_and_baseline_clinical_characteristics.csv
│ │ ├── D02_ROI_level_gene_expression_log2_normalized.csv
│ │ ├── D03_ROI_level_gene_expression_Z_scores_scaled.csv
│ │ ├── D04_patient_level_gene_expression_log2_normalized.csv
│ │ ├── D05_patient_level_gene_expression_Z_scores_scaled.csv
│ │ ├── D06_patient_survival_data_time_to_event_intervals_and_censoring.csv
│ │ ├── D07_gene_level_CoxPH_global_unadjusted.csv
│ │ ├── D08_gene_level_CoxPH_adjusted_for_confounding.csv
│ │ ├── D09_CTA_biological_pathway_annotations_and_gene_sets.csv
│ │ ├── D10_ROI_level_CTA_biological_pathway_activity_Z_scores.csv
│ │ ├── D11_patient_level_CTA_biological_pathway_activity_Z_scores.csv
│ │ ├── D12_CTA_biological_pathway_activity_quality_control_metrics.csv
│ │ ├── D13_patient_level_Bayesian_posteriors_for_CTA_biological_pathways.csv
│ │ ├── D14_Bayesian_CTA_biological_pathways_CoxPH_global_unadjusted.csv
│ │ ├── D15_Bayesian_CTA_biological_pathways_CoxPH_adjusted_for_confounding.csv
│ │ └── D16_per_patient_clinical_notes_deidentified.csv
│ │
│ └── SCAPeSCLC.xlsx
│
├── figures/
│ └── SCAPeSCLC_pipeline.png
│
└── scripts/
├── 01_gene_level_cox_models.R
├── 01_gene_level_cox_ph_assumptions.R
├── 02_bayesian_patient_level_pathways.R
├── 03_pathway_posterior_cox_models.R
├── 03_pathway_posterior_cox_ph_assumptions.R
├── 04_SCAPeSCLC_diagnostic_atlas_generator_for_genes.R
├── 04_SCAPeSCLC_diagnostic_atlas_generator_for_genes_confounder_adjusted.R
├── 05_SCAPeSCLC_diagnostic_atlas_generator_for_BPs.R
└── 05_SCAPeSCLC_diagnostic_atlas_generator_for_BPs_confounder_adjusted.R
The data/ directory contains the three analysis-ready data files directly used as inputs by the R scripts in this repository.
| File | Description |
|---|---|
| D5_scaled_gene_expression.csv | Patient-level standardized gene expression matrix. |
| D10_ROI_CTA_Zscores.csv | ROI-level Cancer Transcriptome Atlas pathway enrichment Z-scores. |
| D13_patient_BP_posteriors.csv | Patient-level Bayesian posterior pathway activity estimates. |
The dataset/ directory contains the complete SCAPeSCLC dataset in two formats:
- CSV format: The
dataset/csv/directory contains the 16 individual data tables comprising the complete dataset. - Excel format:
dataset/SCAPeSCLC.xlsxprovides the complete dataset in a consolidated workbook.
The repository implements the following analytical workflow:
- Bayesian estimation of patient-level pathway activities from ROI-level Cancer Transcriptome Atlas pathway enrichment scores.
- Gene-level Cox proportional hazards regression for progression-free, disease-specific, and overall survival.
- Pathway-level Cox proportional hazards regression using Bayesian posterior pathway activity estimates.
- Assessment of proportional hazards assumptions using Schoenfeld residuals for both gene- and pathway-level Cox regression models.
- Generation of comprehensive diagnostic atlases for both genes and biological pathways, including:
- model summary statistics,
- hazard ratios and 95% confidence intervals,
- Wald test statistics,
- proportional hazards test results,
- Martingale residuals,
- Schoenfeld residuals,
- Deviance residuals,
- DFBETA influence diagnostics.
- Generation of both unadjusted and confounder-adjusted diagnostic atlases for all survival endpoints (OS, DSS, and PFS).
The diagram below summarizes the provenance of the major data products included in SCAPeSCLC and their relationships to the analytical pipelines implemented in this repository.
flowchart TD
A["Public GEO Datasets<br/>GSE261345 & GSE261348"]
A --> B["Clinical and Transcriptomic<br/>Data Harmonization"]
B --> C["D01: Patient Demographics and<br/>Baseline Clinical Characteristics"]
B --> D["D02: ROI-Level Gene Expression<br/>log2-normalized"]
B --> E["D03: ROI-Level Gene Expression<br/>Z-score Scaled"]
B --> F["D04: Patient-Level Gene Expression<br/>log2-normalized"]
B --> G["D05: Patient-Level Gene Expression<br/>Z-score Scaled"]
B --> H["D06: Patient Survival Data<br/>Time-to-Event Intervals and Censoring"]
B --> I["D09: CTA Biological Pathway<br/>Annotations and Gene Sets"]
B --> J["D16: Per-Patient Clinical Notes<br/>(Deidentified)"]
G --> K["Gene-Level CoxPH Regression"]
H --> K
K --> L["D07: Gene-Level CoxPH<br/>Global Unadjusted"]
K --> M["D08: Gene-Level CoxPH<br/>Adjusted for Confounding"]
D --> N["D10: ROI-Level CTA Biological Pathway<br/>Activity Z-scores"]
E --> N
I --> N
N --> O["D12: CTA Biological Pathway Activity<br/>Quality Control Metrics"]
N --> P["Bayesian Patient-Level<br/>Pathway Activity Estimation"]
P --> Q["D13: Patient-Level Bayesian Posteriors<br/>for CTA Biological Pathways"]
Q --> R["Bayesian CTA Biological Pathway<br/>CoxPH Regression"]
H --> R
R --> S["D14: Bayesian Pathway CoxPH<br/>Global Unadjusted"]
R --> T["D15: Bayesian Pathway CoxPH<br/>Adjusted for Confounding"]
K --> U["Gene Diagnostic Atlas Generation"]
U --> V["Gene Diagnostic Atlases<br/>Unadjusted & Confounder-Adjusted"]
Q --> W["Biological Pathway Diagnostic<br/>Atlas Generation"]
W --> X["Biological Pathway Diagnostic Atlases<br/>Unadjusted & Confounder-Adjusted"]
R 4.2 or later is recommended.
Required packages:
install.packages(c(
"dplyr",
"survival",
"broom",
"purrr",
"brms",
"tidyr",
"stringr",
"ggplot2",
"patchwork",
"cowplot",
"gtable"
))If you use SCAPeSCLC in your work, please cite the Zenodo dataset and the accompanying data paper. The GitHub repository provides the analysis code and an accessible copy of the complete dataset.
This project is distributed under the MIT License.