Computational cancer research

OncoForge 0.7.0

OncoForge

An evidence-bound research system for tumor-normal gate hypotheses, explicit normal-tissue exclusions, inspectable provenance, and reproducible computational reports.

Current capabilities

  1. 01Schema-validated evidence fabrics with measured, derived, predicted, inferred, hypothesized, and simulated evidence kept distinct
  2. 02Bounded SINGLE, AND, OR, and AND-NOT gate evaluation against tumor, clone, patient, normal, and critical-normal fields
  3. 03Fail-closed normal-tissue exclusions for active or missing critical-normal evidence
  4. 04Pareto hypothesis selection without a combined truth, safety, or efficacy score
  5. 05Inspectable supporting and contradicting assertions, biomarkers, escape routes, uncertainty, and falsification conditions
  6. 06QSA computation receipts with analytic controls, exact hashes, fallback status, limitations, and verbatim advantage assessment
  7. 07Saved evidence, configuration, report payloads, software commit, algorithm record, and reproducibility hashes
  8. 08A separate conceptual profile and mission simulator for bounded exploratory modeling
Authenticated research workbench

Evidence to a saved, inspectable Target Forge report.

Projects, evidence, runs, and report payloads are account-owned. The portal accepts authorized research datasets, not patient records or clinical instructions.

  1. 01
    Create project

    Keep an account-owned research question, evidence record, and run history together.

  2. 02
    Validate evidence

    Check the OncoForge evidence schema and preserve its origin, counts, classes, categories, and fabric hash.

  3. 03
    Evaluate gates

    Build the tumor-normal model, enumerate the enabled bounded gate families, and apply configured exclusions.

  4. 04
    Inspect normal safety

    Review activation, unknown normal states, missing critical-normal evidence, and every rejection reason directly.

  5. 05
    Read the receipt

    Inspect QSA eligibility or fallback alongside the information-identical classical control and limitations.

  6. 06
    Preserve the run

    Save the exact report, hashes, configuration bounds, and reproducibility fields to the owning account.

Create accountOpen research workspace