Data upload

Oncology studies combine molecular datasets with clinical cohort metadata. Register samples and endpoints in the metadata service before dispatching analysis jobs.

Samples

Register samples via the study Data page or the metadata API:

curl -X POST http://localhost:8000/oncology/studies/{study_id}/samples \
  -H "Content-Type: application/json" \
  -d '{
    "external_id": "S1",
    "patient_id": "P1",
    "timepoint": "baseline",
    "tumor_site": "lung",
    "treatment": "anti-pd1",
    "response_status": "partial_response",
    "group_label": "responder"
  }'

Key fields:

Field	Purpose
external_id	Sample identifier used across pipelines
patient_id	Patient key for clinical endpoint joins
timepoint	Longitudinal position (baseline, on-treatment, relapse)
group_label	Cohort comparison group (responder, non-responder, arm A/B)
response_status	Clinical response category

Clinical endpoints

Add survival and outcome records per patient:

curl -X POST http://localhost:8000/oncology/studies/{study_id}/endpoints \
  -H "Content-Type: application/json" \
  -d '{
    "patient_id": "P1",
    "endpoint_type": "overall_survival",
    "time_to_event_days": 900,
    "event_observed": false,
    "response_status": "partial_response"
  }'

Survival analysis reads clinical CSV files with patient_id, time_to_event_days, and event_observed columns. Endpoint records in the metadata service provide the structured source for export.

Mutation datasets

Register MAF or VCF files linked to the study. The mutation landscape pipeline requires a file path in job parameters:

Parameter	Required	Purpose
mutation_path	Yes	Path to MAF-style CSV on the analysis server
copy_number_path	No	Copy number alteration table
panel_size_mb	No	Sequencing panel size for TMB calculation (default 38.0)
top_genes	No	Number of genes in oncoprint (default 20)

Repertoire datasets

Register TCR/BCR sequencing files for immune profile analysis. Pass the repertoire path in job parameters:

Parameter	Required	Purpose
expression_path	Yes	Bulk or pseudo-bulk expression matrix
repertoire_path	No	TCR/BCR clonotype table

Expression data for communication and immune pipelines

Cell-cell communication requires per-cell expression and metadata CSV files:

File	Required columns
Expression CSV	cell_id plus gene columns
Metadata CSV	cell_id, cell_type; optional condition for comparison

Immune profile requires a bulk expression matrix CSV with sample rows and gene columns.

File storage

Uploads and registered dataset paths live under data/oncology/. The metadata API records file locations in biochem_onco_* tables in Supabase. These paths are gitignored.

Next steps

• File formats — MAF, clinical CSV, expression matrix schemas
• Sample data — test fixtures for local development
• Study workflow