# PACT™ Platform — Specification
## Personalized Antigen-specific Clonal Tolerance
*Tolera Bio's end-to-end engine: from a walk-in patient's biopsy + blood to a bespoke N-of-1 tolerance vaccine.*

---

## 0. One-sentence description
PACT converts a single esophageal biopsy and blood draw into a **patient-specific pMHC-II tolerance vaccine** that retires the food-reactive pathogenic effector-Th2 (peTH2) clone driving that individual's EoE — manufactured on a shared, validated MHC-II backbone into which the patient's own HLA allele and identified food-epitope cassette are dropped.

---

## 1. The N-of-1 principle
Every EoE patient's disease is defined by a personal triad: **{their HLA-II allele} × {their offending food peptide(s)} × {their peTH2 TCR}**. PACT does not guess this triad from a population average — it **reads it out of the individual** and manufactures against it. The shared backbone is what makes bespoke manufacturing tractable: only two variables change per patient (the MHC allele domain and the peptide cassette), so N-of-1 production is a **cassette-swap on a qualified platform**, not a de-novo biologic each time — the same logic that made individualized neoantigen cancer vaccines (mRNA-4157, autogene cevumeran) manufacturable at scale.

---

## 2. End-to-end workflow (7 stages)

| # | Stage | Input | Output | Turnaround (target) |
|---|---|---|---|---|
| 1 | **Sample intake** | Esophageal biopsy + peripheral blood at diagnostic endoscopy | Banked PBMC, tissue, serum | Day 0 |
| 2 | **HLA-II typing** | Blood | Patient's HLA-DRB1/DQ genotype | 2–3 d |
| 3 | **Trigger & clonotype profiling** (the Dx assay) | PBMC + food antigen panel | Ranked food trigger(s) + peTH2 TCR clonotype(s) | 5–10 d |
| 4 | **Epitope core selection** | HLA genotype + trigger protein(s) | Patient-presentable 9-mer core(s) + 15-mer cassette(s) | 1–2 d (compute) |
| 5 | **Construct assembly** | Backbone + patient HLA domain + cassette | Bespoke single-chain pMHC-II sequence(s) | 1 d (compute) → gene synthesis |
| 6 | **Manufacture & formulate** | Construct gene | Expressed pMHC-II → iron-oxide nanoparticle conjugation → release-tested vaccine | 4–8 wk |
| 7 | **Dose & monitor** | Vaccine | Antigen-specific Tr1 induction; peTH2 clone contraction | on-treatment |

**Total vein-to-vaccine target: 6–10 weeks** (dominated by GMP manufacture + release).

---

## 3. The three molecular building blocks (already designed & GPU-validated)
- **Backbone:** shared HLA-DRβ1β2 + DRα1α2 ectodomain single-chain, covalent-peptide topology (L1 = `GGGGSGGGGSGGS`, L2 = `(GGGGS)₄`). Validated on HLA-DRB1*07:01; GPU co-fold ipTM ≥ 0.87, full 15-residue groove engagement.
- **Cassette:** patient-specific 15-mer food epitope, N-terminally tethered into the groove. Dairy anchor (β-casein aa59–78, core `FAQTQSLVY`) is **tetramer-validated**; other food/HLA combinations are computational priors pending wet-lab validation.
- **Multivalent scaffold:** iron-oxide nanoparticle, maleimide surface, free-thiol pMHC conjugation (Navacim/Santamaria precedent) → drives Signal-2 TCR cross-linking and IL-10⁺ Tr1 induction.

---

## 4. Mechanism of action (two signals → tolerance)
1. **Signal-1 (anergy):** the patient's food peptide presented on its restricting MHC-II engages the peTH2 TCR **without full costimulation** → clonal anergy.
2. **Signal-2 (active regulation):** multivalent NP display cross-links the TCR and triggers innate signaling → **antigen-specific Tr1 (IL-10⁺) induction**.
- **Net:** the disease-driving food-reactive clone is silenced and actively regulated; the rest of the T-cell repertoire is untouched — durable, antigen-specific tolerance rather than global immunosuppression.

---

## 5. Personalization routing (see decision-tree figure)
Every patient is manufactured N-of-1, but the **evidence tier** of their cassette determines the development/regulatory route:
- **Tier A — validated cassette** (e.g. DR7 × dairy): patient's components match an already wet-lab-validated backbone+cassette; fastest release, strongest data package.
- **Tier B — computational-prior cassette** (common HLA × wheat/soy/egg): components fall on characterized backbone but the specific epitope is a computational prior → requires the ex-vivo functional-validation gate before dosing.
- **Tier C — rare HLA / novel epitope:** patient's HLA or trigger is outside the qualified set → full N-of-1 characterization workup (new backbone allele modeling + functional validation) before manufacture.

The **product is always personalized**; the platform's job is to move as many patients as possible into Tier A over time by converting validated priors into qualified components.

---

## 6. What makes PACT defensible
- **Validated anchor + computational engine:** one tetramer-validated epitope de-risks the mechanism; the compute pipeline scales it to any HLA×food.
- **Shared-backbone manufacturing:** turns N-of-1 from artisanal into a qualified cassette-swap.
- **Companion diagnostic built in:** Stage 3 is itself a standalone product (names the causal food from blood) and the patient-selection CDx for the vaccine.
- **Founder-patient trial design:** recruitment, endpoints, and community trust.

*All non-dairy binding values are computational priors; PACT is a preclinical-stage design platform, not a validated therapeutic. Not medical advice.*