Candidate Structures
Network Perturbation Map
In silico network linking fibrotic disease state, CRISPR-selected targets, and FAISS-ranked modulators.
Combinatorial Screen Metrics
| Baseline MSE (disease vs blueprint) | 0.0116 |
| Perturbed MSE (after virtual CRISPR) | 0.0118 |
| Phenotypic Reversal Score | 0.0% |
Phenotypic Reversal (%) = max(0, 1 − MSE_perturbed / MSE_baseline) × 100, where MSE is mean squared error vs. fetal regenerative blueprint (lower perturbed MSE = better).
Target Perturbation Profile
Pharmacological Justification StateZero template (no API key)
The StateZero combinatorial perturbation profile (FGF2) implicates pathways controlling epithelial–mesenchymal balance, oxidative stress, and proliferative signaling that distinguish fibrotic burn skin from fetal regenerative blueprint states. The proposed multi-agent regimen (Metformin (CN(C)C(=N)NC(=N)N…); Allopurinol (O=c1[nH]c2ccccc2[nH]c1=O…); Memantine (CC12CC3CC(C)(C1)CC(N)(C3)C2…)) is designed to engage complementary nodes: attenuating pro-fibrotic kinase and mTOR-driven matrix deposition while scavenging reactive oxygen species that sustain myofibroblast activation. Together, these mechanisms plausibly restore a pro-regenerative transcriptional set point aligned with the in silico CRISPR objective.
From a developability perspective, the components occupy drug-like chemical space with favorable synthetic accessibility and separable ADMET liabilities that can be optimized in parallel. Topical delivery with q12h dosing targets sustained exposure within the predicted therapeutic window while limiting systemic burden — an appropriate strategy for cutaneous fibrosis. Experimental validation should prioritize ex vivo human skin models and transcriptomic confirmation that the combination recapitulates the predicted gene expression shift before advancing to in vivo burn wound studies.
Physicochemical Profile (RDKit)
| Property | Metformin | Allopurinol | Memantine |
|---|---|---|---|
| Molecular Weight (< 500 Da) | 129.17 | 162.15 | 179.31 |
| LogP (< 5) | -1.03 | 0.22 | 2.69 |
| TPSA (< 140 Ų) | 88.99 | 65.72 | 26.02 |
| H-Bond Donors (≤ 5) | 4 | 2 | 1 |
| H-Bond Acceptors (≤ 10) | 2 | 2 | 1 |
| Lipinski Rule of 5 | Pass | Pass | Pass |
ADMET & Toxicity Profile (PyTDC)
PyTDC ML oracles unavailable — showing RDKit physicochemical and PK/PD proxy scores.
| Endpoint | Metformin | Allopurinol | Memantine |
|---|---|---|---|
| Drug-likeness (QED) (0–1) | 0.249 | 0.544 | 0.608 |
| Synthetic accessibility (1–10) | N/A | N/A | N/A |
| LogP | -1.034 | 0.216 | 2.694 |
| DRD2 activity (proxy) (prob.) | N/A | N/A | N/A |
| GSK3β activity (proxy) (prob.) | N/A | N/A | N/A |
| JNK3 activity (proxy) (prob.) | N/A | N/A | N/A |
| CYP3A4 inhibition (prob.) | N/A | N/A | N/A |
| Predicted clearance (L·h⁻¹) | 2.678 | 1.940 | 1.781 |
| Volume of distribution (L) | 113.597 | 164.629 | 179.872 |
| Elimination rate k_e (h⁻¹) | 0.024 | 0.020 | 0.020 |
| Molecular weight (g·mol⁻¹) | 129.167 | 162.148 | 179.307 |
Pharmacokinetics — Topical 1-Compartment Model
Hybrid AI–mechanistic simulation (q12h topical dosing). Solid lines = each FAISS-ranked component; dashed red = combination envelope (per-timepoint max). Green band = therapeutic window (MEC–MTD).
