Generate publication-ready Gene Set Enrichment Analysis (GSEA) visualizations:
- enrichment score (ES) curves;
- ranked-gene (RNK) panels; and
- leading-edge (LE) expression heatmaps.
| Item | Location |
|---|---|
| Canonical module | OMIX GSEA Visualization Legacy |
| Interface contract | schemas/interface.yml |
| Development contract | OMIX module contract |
| Released source reference | OMIX_MODULE_SOURCE.md |
The canonical module owns scientific behavior, portable CLI operation, tests, and the reusable input/output contract. This repository is its Code Ocean deployment adapter.
- The Code Ocean App Panel and capsule entry point.
- Recursive workflow-result discovery and explicit DEG/metadata overrides.
- The input bundle and result layout used by the OMIX DEG-to-GSEA workflow.
The capsule expects three inputs:
- MSigDB database — the same MSigDB release used for upstream GSEA.
- Filtered GSEA results —
filtered_gsea_results.csvfrom OMIX GSEA Filters. - DEG Analysis result bundle — one OMIX DEG Analysis Result containing
both
DEG_Analysis.csvandSample_Metadata.csvin the same folder.
The DEG table supplies the ranking statistics and sample-level expression. If batch adjustment was included in DEG Analysis, its appended expression columns are batch-adjusted voom-scale log-CPM values and are used directly for the LE heatmap. The matching metadata controls sample annotation and ordering.
Keep DEG_Analysis.csv and Sample_Metadata.csv together. The adapter stops
on ambiguous or mismatched bundles instead of silently using files from
different analyses.
For an ad-hoc override, users may optionally provide both a DEG table and its matching sample-metadata table in the app panel. The two explicit files take precedence over the attached bundle and cannot be supplied independently.
Attach the three inputs above, then choose the pathway selection and display
settings in the app panel. Workflow Results may mount in generated
subdirectories below /data; the adapter discovers the required files
recursively.
The input bundle may also contain the DEG run summary and diagnostic images. Those provenance files are preserved but ignored by the visualization step.
The equivalent command-line interface is:
Rscript code/main.R \
--msigdb_database /path/to/MSigDB_v2023_2.rds \
--gsea_filter_results /path/to/filtered_gsea_results.csv \
--deg_analysis_results /path/to/deg-analysis-result \
--plots_to_include ES+RNK+LE \
--output_dir results--deg_analysis_results must be the directory containing both portable DEG
output tables, not just DEG_Analysis.csv alone.
Alternatively, provide both --deg_table and --sample_metadata to override
the bundle with an explicit matched pair.
GSEA-Vis-Enrichment-Plots.pdf— selected pathway plots in a multi-page PDF.GSEA-Vis-RunningES.csv— ranked-gene and running-enrichment-score values.
Run logs also report input-consistency checks, including whether the filtered GSEA pathways and DEG ranking statistics are compatible.
Use the pinned capsule environment defined in environment/. Retain the
MSigDB release, filtered-GSEA result identity, DEG-result bundle, and selected
plot parameters with every released figure.
code/
main.R Code Ocean and command-line entry point
functions/gsea_enrichment_plot.R Preserved visualization implementation
README.md Detailed adapter documentation
.codeocean/ App-panel and default-data configuration
environment/ Capsule Docker environment
“Legacy” identifies the established visualization implementation this adapter preserves; it does not change the required input validation or provenance checks.
Read AGENTS.md and OMIX_MODULE_SOURCE.md before editing. Reusable scientific changes belong in the canonical OMIX module and are exported here only after validation.