#!/usr/bin/env python3 """ comp-001: Uricase protease stability in shio-koji fermentation Orchestrator — loads inputs, calls shared library, writes comp-001-specific outputs. Core algorithm lives in experiments/lib/protease_stability.py. Usage: python3 analyze.py (from this directory) Outputs: outputs/cleavage_sites.json, outputs/summary.md """ import json import sys from pathlib import Path sys.path.insert(0, str(Path(__file__).parent.parent / "lib")) from protease_stability import ( load_sequence, load_plddt, load_proteases, compute_sequence_stats, run_all_proteases, ) INPUTS = Path(__file__).parent / "inputs" OUTPUTS = Path(__file__).parent / "outputs" OUTPUTS.mkdir(exist_ok=True) def main(): seq = load_sequence(INPUTS / "Q00511.fasta") plddt = load_plddt(INPUTS / "alphafold_Q00511_plddt.json") proteases, conditions = load_proteases(INPUTS / "protease_specificities.json") sequence_stats = compute_sequence_stats(plddt) protease_results = run_all_proteases(seq, plddt, proteases, conditions) output = { "protein": "Uricase (urate oxidase), Aspergillus flavus (Q00511)", "conditions": conditions, "sequence_stats": sequence_stats, "protease_results": protease_results, } with open(OUTPUTS / "cleavage_sites.json", "w") as f: json.dump(output, f, indent=2) write_summary(output, OUTPUTS / "summary.md") print("Done. Outputs written to outputs/") def write_summary(data, path): ss = data["sequence_stats"] cond = data["conditions"] pr = data["protease_results"] lines = [ "# comp-001 — Uricase Shio-Koji Protease Stability: Summary", "", f"**Protein:** {data['protein']} ", f"**Conditions modeled:** {cond['NaCl_pct']}% NaCl, pH {cond['pH_range'][0]}–{cond['pH_range'][1]}, {cond['temperature_C']}°C, {cond['duration_days'][0]}–{cond['duration_days'][1]} days ", f"**Analysis date:** 2026-05-05 ", f"**Script:** `experiments/comp-001-uricase-shio-koji-protease-stability/analyze.py` ", "", "---", "", "## Structural Overview", "", "| Metric | Value |", "|---|---|", f"| Sequence length | {ss['length']} aa |", f"| Mean pLDDT (AlphaFold confidence) | {ss['mean_plddt']} / 100 |", f"| Minimum pLDDT (most flexible residue) | {ss['min_plddt']} / 100 |", f"| % residues pLDDT > 80 (well-folded) | {ss['pct_residues_above_80']}% |", f"| % residues pLDDT > 90 (core-folded) | {ss['pct_residues_above_90']}% |", "", "**Interpretation:** A mean pLDDT of 97.1 with a minimum of 80.5 is exceptional — essentially the entire protein is predicted to be in a well-folded conformation with no disordered loops or flexible termini. This is the primary structural argument for protease resistance: there are no exposed unstructured regions for proteases to initiate cleavage.", "", "---", "", "## Per-Protease Risk Assessment", "", ] for name, r in pr.items(): lines += [ f"### {r['full_name']} (`{name}`)", "", "| Parameter | Value |", "|---|---|", f"| Recognition sites in sequence | {r['total_recognition_sites']} |", f"| Buried (pLDDT ≥ 80) | {r['buried_sites']} |", f"| Partially exposed (pLDDT 65–80) | {r['partially_exposed_sites']} |", f"| Exposed (pLDDT < 65) | {r['exposed_sites']} |", f"| Residual activity at 17.5% NaCl | {r['salt_residual_activity_at_nacl']*100:.0f}% |", f"| pH activity factor (shio-koji pH) | {r['ph_activity_factor']*100:.0f}% |", f"| Effective protease activity (salt × pH) | {r['effective_activity']*100:.1f}% |", f"| Max risk score (0–1) | {r['max_risk_score']} |", "", ] if r["top_5_sites"]: lines += [ "**Top 5 highest-risk cleavage sites:**", "", "| Position | P1 | P1' | pLDDT (window) | Accessibility | Risk score |", "|---|---|---|---|---|---|", ] for s in r["top_5_sites"]: lines.append( f"| {s['position']} | {s['P1']} | {s['P1_prime']} | " f"{s['mean_plddt_window']} | {s['accessibility']} | {s['risk_score']} |" ) lines.append("") lines += [ "---", "", "## Overall Risk Verdict", "", ] max_risks = {n: r["max_risk_score"] for n, r in pr.items()} worst = max(max_risks, key=max_risks.get) worst_score = max_risks[worst] if worst_score < 0.15: verdict = "**LOW** — protease degradation in shio-koji is unlikely to meaningfully reduce uricase activity" color = "Low" elif worst_score < 0.30: verdict = "**MODERATE** — some degradation possible; wet-lab confirmation (§1.10) recommended before relying on shio-koji format" color = "Moderate" else: verdict = "**HIGH** — significant protease risk; shio-koji format likely incompatible with uricase delivery" color = "High" lines += [ f"**Overall risk: {color}**", "", verdict, "", f"The highest single-site risk score is {worst_score} (from `{worst}` — {pr[worst]['full_name']}). ", "", "### Key factors driving the verdict", "", "1. **Exceptional structural stability (pLDDT 97.1 mean, 80.5 minimum).** All potential cleavage sites in the sequence are located in confidently-folded regions. There are no disordered loops or exposed termini — the primary entry points for protease attack.", "2. **Strong ALP salt inhibition.** ALP (the dominant koji protease) retains only ~10–20% activity at 15–20% NaCl. Shio-koji's defining salt level substantially neutralises the most active protease.", "3. **Uricase is a homotetramer.** This analysis models the monomer. In the native quaternary structure, subunit interfaces bury additional surface area, further protecting internal residues from protease access. The monomer analysis is conservative.", "4. **Acid protease is active at shio-koji pH (~4.5–5.0) but at reduced efficacy (~30% of pH-optimal).** It is also the most salt-tolerant of the three — the residual risk from the acid protease is the most meaningful of the three, but still modest given that all its recognition sites are in folded regions.", "", "### Limitations", "", "- pLDDT ≠ solvent accessibility. Some high-pLDDT residues on protein surface loops may still be accessible. A molecular dynamics simulation or explicit solvent-accessibility calculation would sharpen this.", "- P1/P1' rules only. Real protease extended binding subsites (P2–P4) are not modelled; this may over-count recognition sites.", "- Monomer structure only. Quaternary burial (tetramer interfaces) would reduce accessible surface further — this analysis is conservative.", "- ALP and NPr pH factors set to 1.0 (conservative). ALP is outside its active pH range at shio-koji pH 4.5–5.0; NPr is at its lower edge. True risk from these two proteases is lower than computed.", "- No fermentation dynamics. During active koji growth (before shio-koji is made), proteases operate at higher activity. The shio-koji format specifically starts after koji is harvested and mixed with salt — the peak-activity phase is before the salt environment.", "", "### Recommended action", "", "The structural analysis supports the hypothesis that uricase will be substantially resistant to shio-koji proteolysis. **`validation-experiments.md` §1.10 should still be run** — the experiment remains the ground truth — but this analysis shifts the prior from 'unknown' to 'probably fine', which changes the priority framing: §1.10 is a confirmation experiment, not a make-or-break feasibility gate.", "", "If §1.10 shows unexpected degradation, the primary suspects are: (a) the acid protease operating at pH 4.5–5.0 on any surface-accessible sites; (b) cooperative unfolding of the monomer in the salt/acid environment (the structure is stable in silico under standard conditions; shio-koji conditions may differ). A follow-up comp-002 could model thermal/pH stability explicitly.", "", "---", "", "*Generated by `analyze.py` on 2026-05-05. Re-run after any AlphaFold model update or MEROPS specificity revision. See `inputs/provenance.md` for data sources.*", ] with open(path, "w") as f: f.write("\n".join(lines) + "\n") if __name__ == "__main__": main()