AD Quantum Repurposing (Proposal 1 Baseline)

Goal: classical docking + BBB heuristics + quantum demo to scaffold a quantum-enhanced repurposing pipeline for Alzheimer’s disease (AD).

What’s in this repo

• data/raw/ – PDB 4DVF and seed SMILES
• data/processed/ – prepared receptor & ligands (PDBQT/SDF)
• results/docking/ – docking logs/poses + docking_results.csv
• results/reports/ – ranking.md, ranking_table.csv, docking_bar.png, vqe_demo.md
• results/quantum/ – vqe_h2.json
• scripts/ – helper scripts (incl. 06_vqe_workflow.py)
• envs/wsl_adquant_q.yml – WSL environment for Qiskit/PySCF

Methods (summary)

• Protein prep: PDBFixer/OpenMM on 4DVF; box center [3.061, −0.128, 17.532], size [22,22,22] Å.
• Ligands: 10 CNS-approved drugs; RDKit for 3D + MW/PSA/cLogP; OpenBabel for PDBQT & Gasteiger charges.
• Docking: smina (Vina score); parsed best-pose energies; composite rank = binding + BBB features.
• Quantum demo: Qiskit Nature VQE (H₂) to validate the quantum stack; outputs JSON + report.

Results (top-10)

Rank	Ligand	Affinity (kcal/mol)	MW	PSA	cLogP	BBB	Composite
1	donepezil	−8.2	392.21	42.01	3.856	✅	2.2470
2	galantamine	−7.6	278.09	39.44	4.245	✅	1.5164
3	nilvadipine	−7.0	433.05	78.51	4.464	✅	0.7858
4	fluoxetine	−6.7	259.14	35.25	3.796	✅	0.4205
5	sertraline	−6.6	283.09	12.03	4.467	✅	0.2987
6	riluzole	−6.4	253.02	76.72	2.140	✅	0.0552
7	rivastigmine	−6.1	221.11	38.77	2.036	✅	−0.3101
8	memantine	−5.9	181.18	26.02	3.084	✅	−0.5537
9	rasagiline	−5.7	191.11	12.03	2.175	✅	−0.7972
10	selegiline	−5.4	173.12	12.03	2.036	✅	−1.1625

Highlights

• 10/10 ligands docked and parsed (100% success).
• Best vs worst |ΔE| improvement: ~52% (donepezil vs selegiline).
• Best vs median |ΔE| improvement: ~26% (donepezil vs median).
• All pass BBB heuristics used here (CNS-plausible).

See results/reports/ranking.md and results/reports/docking_bar.png.

Quantum check

VQE demo for H₂ succeeded (see results/quantum/vqe_h2.json, results/reports/vqe_demo.md). This validates Qiskit/PySCF on WSL and is a placeholder for target-specific quantum simulations we’ll add next.

Limitations

Rigid receptor; approximate scoring; single target; simple composite score; no experimental validation yet.

Next steps

• Rescoring (RFScore/MMGBSA), interaction analysis, ensemble docking.
• Add a second AD target (e.g., tau) for polypharmacology.
• Swap demo H₂ for ligand-fragment active-space VQE.
• Plan organoid/iPSC validation with ≥30% reduction thresholds for Aβ/p-tau.