Research & Paper
Results as protocol: write the reproducible path first.
The draft emphasizes methods, protocols and figure placeholders. Submission status: in preparation. DOI TBD. Fit errors, measured coverage and pass/fail conclusions await public-data runs.
§ 1
Paper status
Primary readers: research / QSP reviewers. The sell is a reviewable evidence chain — not flashier animation.
§ 2
Methods summary
Hybrid core (ABM + ODE/diff + 2D RD); scenario packs; explicit measurement models; calib/validation isolation; UQ with coverage + sharpness; L1/L2; Verification-lite; reports cite replay.
- Six-gap countermeasures align with paper §1
- Murine public data for baseline
- AI strategy search excludes molecule generation
§ 3
Reproduction protocol skeleton
Structure example for reviewer checklists.
reproduction_checklist.md
1. Pin commit / scenario_version / param_set_id
2. Declare train vs validate task files
3. Calibrate → calibration_report.json
4. UQ → uq_report.json (incl. sharpness)
5. L1 (optional L2) → validation_report.json
6. Verification-lite → verify_dt_grid_*.yaml
7. Pack artifact.zip (incl. replay index)§ 4
Figure checklist (placeholders)
| Item | Intent | Status |
|---|---|---|
| Fig. mech | Mechanism schematic | Schematic |
| Fig. calib | Calibration alignment (public data) | Pending public-data runs |
| Fig. uq | Interval coverage + sharpness | Pending public-data runs |
| Fig. L1/L2 | Extrapolation tasks & outcomes | Pending public-data runs |
| Table thresholds | Default thresholds vs COU overrides | Document targets |
§ 5
Data Availability template
data_availability.txt
Public data sources: <citation or URL placeholders>
Digitization scripts & hashes: <paths>
Scenario pack version: <scenario_version>
Code repo & commit: <repo>@<commit>
Limits: individual clinical data out of v1 scope§ 6
Default scientific spine
Default: murine Influenza-like respiratory virus + MC38-like tumor + pharmacologic blockade/exhaustion L2. Replaceable with explicit COU and data-list updates.