Wang TNBC spatial transcriptomics processing

All scripts are stored directly in the ST_TNBC_Wang project folder.

Folder layout

ST_TNBC_Wang/
├── 00_wang_config.R
├── 01_wang_gene_sets.R
├── 01_wang_preprocess_functions.R
├── 02_run_wang_full_cohort.R
├── 03_validate_wang_marker_scores.R
├── 04_prepare_wang_release.R
├── Clinical/
├── classification/
├── Robjects/
└── outputs_wang/

First run: pilot

In 00_wang_config.R, use:

start_parent = 1
stop_parent = 3
overwrite = TRUE

Run:

Rscript 02_run_wang_full_cohort.R

Review the pilot outputs.

Full corrected cohort

Delete the pilot outputs_wang folder or keep overwrite = TRUE.

Change:

start_parent = NULL
stop_parent = NULL

Run:

Rscript 02_run_wang_full_cohort.R

The preprocessing step directly creates:

• normalized expression matrices;
• spot tables in RDS and CSV.GZ formats;
• Seurat objects;
• QC summaries;
• module-coverage summaries.

Marker-score validation

Run:

Rscript 03_validate_wang_marker_scores.R

This validates tumor, broad lymphocyte, and TLS candidate scores against manual annotations.

Lineage-specific T-cell, CD8/cytotoxic, B-cell, plasma-cell, macrophage, and M2 variables remain marker-module enrichment scores, not deconvolved proportions.

Prepare documentation

Run:

Rscript 04_prepare_wang_release.R

This does not recreate the biological outputs. It creates manifests, checksums, the data dictionary, module definitions, and release documentation inside outputs_wang.

GitHub versus Dropbox

Private GitHub repository:

• configuration template;
• gene sets;
• all R scripts;
• this README.

Dropbox:

• upload the complete verified outputs_wang folder.