Mutation to outcomes

Connect somatic mutation landscapes to clinical endpoints — TMB, driver alterations, mutational signatures, and oncoprint summaries stratified by survival or response.

Research question

Which mutational features — TMB, specific driver genes, copy number alterations, or mutational signatures — associate with overall survival, progression-free survival, or treatment response in my cohort?

Who this is for

• Translational oncology teams at cancer centers and precision medicine programs
• Computational oncology groups processing MAF cohorts with clinical follow-up
• Pharma biomarker teams linking sequencing endpoints to trial outcomes

Data requirements

Data	Required	Purpose
MAF-style mutation table	Yes	Variant classification, TMB, oncoprint, signatures
Clinical endpoint table	Yes (for survival)	Kaplan-Meier, log-rank, Cox regression
Copy number alteration table	No	Amplification/deletion landscape
Molecular feature matrix	No	Multivariate Cox with TMB + other features

Align Tumor_Sample_Barcode in the MAF with patient_id in clinical records via sample metadata.

Workflow

Register samples and clinical endpoints
  → Mutation landscape (TMB, oncoprint, signatures)
  → Survival analysis stratified by TMB or driver status
  → AI interpretation

Step 1 — Mutation landscape

Launch with cohort MAF and optional copy number file:

curl -X POST http://localhost:8001/oncology/jobs/mutation-landscape \
  -H "Content-Type: application/json" \
  -d '{
    "study_id": "your-study-id",
    "parameters": {
      "mutation_path": "/path/to/cohort.maf",
      "copy_number_path": "/path/to/copy_number.csv",
      "panel_size_mb": 38.0,
      "top_genes": 20
    }
  }'

Review per-sample TMB, gene frequency table, oncoprint matrix, co-occurrence/exclusivity pairs, and dominant mutational signature on the Mutations page.

Step 2 — Survival stratification

Export TMB or driver mutation status as features, then run survival analysis:

curl -X POST http://localhost:8001/oncology/jobs/survival \
  -H "Content-Type: application/json" \
  -d '{
    "study_id": "your-study-id",
    "parameters": {
      "clinical_path": "/path/to/clinical.csv",
      "feature_path": "/path/to/tmb_features.csv",
      "stratify_by": "treatment_arm",
      "cox_covariates": ["tmb"]
    }
  }'

Review Kaplan-Meier curves, log-rank p-values, and Cox hazard ratios on the Survival page.

Step 3 — Interpretation

Include both mutation landscape and survival run IDs in an interpret job for a multi-modal narrative linking alterations to outcomes.

Expected outputs

• Per-sample TMB and variant classification breakdown
• Oncoprint matrix with clinical annotation tracks
• Mutational signature decomposition (COSMIC SBS/DBS/ID)
• Driver gene and pathway enrichment across subgroups
• Kaplan-Meier curves stratified by TMB quartile, driver status, or treatment arm
• Cox regression with hazard ratios for molecular covariates

Typical analyses

Analysis	Stratification variable	Question
TMB and IO response	TMB high vs. low	Does mutation burden predict checkpoint inhibitor benefit?
Driver co-occurrence	TP53 + KRAS vs. either alone	Are alteration patterns mutually exclusive or co-occurring?
Signature and survival	APOBEC-dominant vs. other	Does mutational process associate with outcome?
Treatment arm comparison	anti-PD-1 vs. chemotherapy	Do molecular features differ in prognostic value by regimen?

Related guides

• Mutation landscape
• Survival analysis
• Sample data — MAF and clinical test fixtures