QAOA · De novo sequence optimizer
QC · 6q × p=2 · H → [RZ(2γ) · CX-RZ(2β)-CX · RX(2β)] × p
ΔG_bind = ⟨ψ(θ)|Ĥ_elec|ψ(θ)⟩ − TΔS · Ising cost over residue-residue distance matrix
VQE output · binding & pharmacology
ΔG_bind (Ha)
-79.670
Affinity K_d
2446.22 nM
Immunogenicity idx
13.2
Half-life
58.2 h
Circuit depth
54
Ansatz
UCC(θ)
Scientific validation guardrails
BLOCK · do not release- ΔG_bind physical range — -79.67 Ha vs. bounded window [−140, −20]FDA-CDER 2021 · Rosetta ref-set
- Immunogenicity index ≤ 12 — IEDB net-MHCII score 13.2IEDB · Prot-MHCII v4.3
- Kd < 50 nM for mint — Kd = 2446.22 nMBiotx-QC v2
- Half-life ≥ 24h — t½ = 58.2 hPK-guardrail
L6 Directive: 🛑 BLOCKED · guardrail violation — do not mint. Re-parameterize ansatz and re-run VQE.
Batch Quantum Simulation · VQE/QAOA sweep
Parallel ΔG_bind distributions across N targets · shot-noise via COBYLA variance · constraint heatmap over guardrails
Target library (4/6)
σ(ΔG) ≈ 0.9 + 1.6/√shots + 0.12·(6−p) · noiseless statevector + finite-sample COBYLA
Guardrail-gated release
Every output blocked or flagged if it violates ΔG / MPa / LoD / Kd constraints
Quantum sim mode
VQE/QAOA convergence traces on ΔG_bind with target-vs-trajectory plots
Guided wet-lab
5-step TX-TL wizard with sample ID, operator sign-off, and softmax over 6 classes
Signed PDF reports
One-click export · run metadata, inputs, outputs, guardrails, evidence links