Pipelines
All oncology pipelines run in backend-analysis and are dispatched asynchronously. List live definitions and parameter schemas:
curl http://localhost:8001/oncology/pipelines
Cell-cell communication
| Pipeline | Input | Outputs |
|---|---|---|
cell_communication | Per-cell expression CSV + metadata CSV | Interaction scores, pathway aggregation, network, ligand activity, condition comparison |
Key parameters: expression_path, metadata_path, condition_column, condition_a, condition_b, permutations.
Reference algorithms: CellPhoneDB, NicheNet, CellChat.
Immuno-oncology profiling
| Pipeline | Input | Outputs |
|---|---|---|
immune_profile | Bulk expression CSV (+ optional repertoire CSV) | Deconvolution, TIDE phenotype, exhaustion scores, repertoire diversity, IO response prediction |
Key parameters: expression_path, repertoire_path.
Reference algorithms: CIBERSORTx, xCell, TIDE, scirpy.
Mutation landscape
| Pipeline | Input | Outputs |
|---|---|---|
mutation_landscape | MAF-style CSV (+ optional copy number CSV) | TMB, oncoprint, signatures, co-occurrence/exclusivity, copy number, pathway enrichment |
Key parameters: mutation_path, copy_number_path, panel_size_mb, top_genes.
Reference algorithms: maftools, SigProfilerPy.
Survival analysis
| Pipeline | Input | Outputs |
|---|---|---|
survival | Clinical CSV (+ optional feature and longitudinal CSVs) | Kaplan-Meier curves, log-rank test, Cox regression, longitudinal trajectories |
Key parameters: clinical_path, feature_path, stratify_by, cox_covariates.
Library: lifelines.
Interpretation
| Pipeline | Input | Outputs |
|---|---|---|
interpret | Completed source run IDs | Narrative JSON with metric citations and research disclaimer |
Key parameters: source_run_ids, interpretation_type.
Shared with Computational Biology
These capabilities are not duplicated in oncology pipelines:
| Capability | Owner |
|---|---|
| Single-cell QC, clustering, DE | Computational Biology |
| Spatial transcriptomics ingestion | Computational Biology |
| Biomarker ML and pathway enrichment | Computational Biology |
Oncology TME views call into compbio infrastructure. See Tumor microenvironment.
Job lifecycle
queued → running → complete | failed
Poll status: GET /oncology/jobs/{run_id}
Job dispatch endpoints:
POST /oncology/jobs/cell-communication
POST /oncology/jobs/immune-profile
POST /oncology/jobs/mutation-landscape
POST /oncology/jobs/survival
POST /oncology/jobs/interpret
Versioning
Every run records pipeline_version (currently 0.1.0). Include run IDs and versions in methods sections for reproducibility.
Implementation sources
Oncology-specific algorithms are ported from reference implementations under products/oncology/applications/. See products/oncology/product.md and products/oncology/todo.md for implementation status.