NEW · Versioned model · 2026-07-07T13:08:18Z

AIG‑VHT Autoimmune Gastritis QSP v1.1

Executable AI‑QSP model of autoimmune gastritis linking parietal-cell loss, acid feedback, gastrin, ECL/CgA, B12/iron and autoantibody dynamics. Added to ATLAS as aig_vht_v1_1_core_calibrated_20260707_1308Z.

CORE_CALIBRATED

Model description

The AIG‑VHT model is a Tier‑1, cross‑sectional calibrated model for steady‑state biomarker interpretation in autoimmune gastritis. It couples the autoimmune corpus module with the acid → gastrin → ECL trophic loop and intrinsic-factor gated B12/iron outputs.

  • Eight dynamic states: parietal mass, ECL mass, Teff, Treg, autoantibody, gastrin, B12 and iron.
  • Calibrated to digitised literature cohort means for gastrin, chromogranin A and vitamin B12.
  • Profile-likelihood identifiability analysis for acid-feedback and ECL parameters.
  • Includes SBML, OMEX archive, calibration targets, calibration result, profile result and manuscript files.

Use and limitations

Status: preprint / medRxiv; not peer reviewed. The calibrated biomarker arm is suitable for research interpretation and experimental-design prioritisation only.

Can do: steady-state biomarker cascade interpretation, identifiability analysis, N-of-1 overlay, experimental agenda design.

Cannot do: clinical diagnosis, treatment selection, patient-specific prognosis or validated digital-twin use. Therapy scenarios are hypothesis-generating.

Mechanistic structure

The model schematic links corpus autoimmunity to circulation readouts and therapeutic levers including low-dose IL‑2, antigen-specific tolerance, netazepide, CAAR‑T and parenteral B12/iron.

AIG‑VHT model schematic
AIG‑VHT model schematic from the uploaded manuscript.
AIG‑VHT calibration and profile likelihood figure
Calibration and identifiability outputs: gastrin, CgA, B12, APCA overlay and profile likelihoods.

Free vs subscription boundary

ATLAS Open

  • Public model card and status labels
  • SBML v1.1 CORE_CALIBRATED branch
  • Calibration target CSV and summary JSON
  • Manuscript/preprint files and figures