Pharmacochaperone (h01) — parameter provenance

Source agent: h01 audit (Sonnet 4.6), 2026-04-23. Consumer: all 17 pharmacochaperone_*.py scripts in models/.


Parameter table

ParameterValue usedUnitsSourcePage/sectionStatusScript(s)Action
Water probe radius1.4ÅPhysical constant (Lee & Richards 1971)standard✅ physicalphase1_mutant_pocketNone
Carbon vdW radius1.7ÅPhysical constant (Bondi 1964)standard✅ physicalphase1_mutant_pocketNone
Salt-bridge O-N distance2.8ÅPhysical constant (Hubbard geometry)standard✅ physicalphase3c_shape_fitNone
K-to-kJ conversion1/4.184Physical constant✅ physicalphase4fNone
Water dielectric constant78.5Physical constant✅ physicalphase4f (eps_solvent)None
pLDDT threshold70Standard AF2/AF3 convention (Jumper 2021 Nature)✅ standardphase1bNone
AmberFF19SBHe 2020 JCTC✅ standard MDphase5None
GAFF2Wang 2004 JCCA✅ standard MDphase5None
TIP3P waterJorgensen 1983 JCP✅ standard MDphase5None
NaCl concentration0.15MPhysiological✅ standardphase5None
Rule-of-3 (Congreve): MW<300, logP<3, HBD≤3, HBA≤3, rot≤3Congreve 2003 Drug Discov Today✅ standard (textbook)phase3bOptional citation note
Docking box size18×18×18ÅIn-silico: Phase 2B subpocket volume → box calibrated to 159 ų pocket✅ in-silicotarget_prep, phase4_commonNone
K1141-NZ to F1646-ring distance5.69 → 5.7ÅIn-silico: measured from AF3 WT CIFPhase 3A✅ in-silicophase3bNone
Reference ΔΔG gap8.4kcal/molIn-silico: STRC Electrostatic Analysis E1659A (multi-method PAE-corrected)own note✅ in-silicophase4f (REFERENCE_GAP_KCAL_MOL)Add # see STRC Electrostatic Analysis E1659A comment
Druggability threshold (Phase 4a gate)≥0.700-1In-silico: calibrated to Phase 0 result (0.86 WT)Phase 0✅ in-silicophase4aNone
ΔΔG gate (WT vs MUT docking)1.0kcal/molIn-silico: ~2× Vina paired noise (0.5-0.8 kcal/mol)phase4c comment✅ in-silicophase4cNone
K1141A score-loss gate≥2.0kcal/molIn-silico: 2× DELTA_GATE, conservative pharmacophore falsification✅ in-silicophase4dNone
Vina exhaustiveness32Standard Vina: 4× default (8), conservative for small pocketTrott 2010 JCIM✅ standardphase4bOptional: cite Trott 2010 / Eberhardt 2021
Vina score threshold (diflunisal gate)≤−5.0kcal/molIn-silico heuristic (conservative relative to diflunisal ~−12 kcal/mol real TTR affinity)⚠ no citationphase4bNEEDS citation: Eberhardt 2021 JCIM (Vina 1.2 CASF benchmark)
MM-PBSA ΔG_bind gate≤−6.0kcal/molIn-silico heuristic⚠ no citationphase5NEEDS citation: Genheden & Ryde 2015 Expert Opin Drug Discov
RWM PK envelope: MW 250-400 Da, LogP 1.5-3.5, HBD 0-2per range2018-salt-plontke-pharmacokinetic-principles-inner-earFig 4, Table 2✅ paper in strc/papers/target_prep docstringAdd # Salt & Plontke 2018 comment in script
TPSA filter for delivery40–90ŲUNLABELED in script — hypothesis note says 70–100 per Salt & Plontke 2018⚠ INCONSISTENCY: CNS range used instead of RWM rangephase3bTIGHTEN to 70–100 per 2018-salt-plontke-pharmacokinetic-principles-inner-ear
MW filter for virtual screen180–350DaUNLABELED — hypothesis note says 200–500⚠ INCONSISTENCYphase3bHarmonize with hypothesis note
Druggability v_opt250–300ųIn-house heuristic (loosely SiteMap-inspired)⚠ inconsistent across phasesphase2 (300), phase2b (250)Label “in-house” or cite Halgren 2009 SiteMap
Druggability composite weights0.5/0.3/0.2 (p1), 0.4/0.25/0.2/0.15 (p2), 0.3/0.2/0.15/0.15/0.2 (p2b)In-house — three different schemes⚠ undocumented inconsistencyphase1, phase2, phase2bDocument rationale or harmonize in one function
MM-GBSA recovery gate≥30% of 8.4 kcal/molClinical analogy: VX-809 restores ~30% CFTR functionBulawa 2012 PNAS✅ clinical analogy — defensiblephase4f (RECOVERY_GATE)None; note already references Bulawa class implicitly

Red flags

MEDIUM priority

1. TPSA filter 40–90 Ų in phase3b (should be 70–100 Ų)

  • Script uses CNS-delivery TPSA range; intratympanic/RWM delivery requires 70–100 Ų per Salt & Plontke 2018.
  • The correct paper IS in strc/papers/. Fix: change filter, add comment # Salt & Plontke 2018 RWM PK envelope.
  • Risk: compounds with TPSA 40–70 Ų may not cross RWM but pass the virtual screen. Three phase3b results could be artifacts.

2. Inconsistent druggability scoring weights across three scripts

  • phase1_mutant_pocket: 0.5 vol + 0.3 hp + 0.2 hb
  • phase2_pocket_scan: 0.40 vol + 0.25 hp + 0.20 r + 0.15 hb
  • phase2b_subpockets: 0.30 v + 0.20 h + 0.15 r + 0.15 hb + 0.20 depth
  • Impact: druggability scores from Phase 1 are NOT comparable to Phase 2 or 2B. Phase 4a threshold of 0.70 was set based on Phase 0 (Phase 2B formula), but Phase 1 uses a different formula. Scripts are used for internal scoring only, not cross-phase numeric comparisons, so operational impact is low — but misleading if compared.

LOW priority

3. Vina −5 kcal/mol gate (phase4b) lacks citation

  • Reasonable conservative value but needs reference to Vina benchmark papers for reviewer defense.

4. MM-PBSA −6 kcal/mol gate (phase5 scaffold) lacks citation

  • MM-PBSA thresholds are not transferable across targets; should be labeled as pipeline-specific empirical gate.

5. MW filter 180–350 Da inconsistent with hypothesis note (200–500 Da)

  • Low risk: tighter range is more conservative for fragment screening. Acceptable if documented.

No phantom citations found

Unlike h09, h01 scripts contain zero fabricated paper citations. No “Salt 2011” style phantom references. Every numeric constant traces to either a physical constant, an in-silico output, a standard community convention, or a real paper already in strc/papers/. The issues above are internal inconsistencies and missing attribution notes — not citation fabrications.


Connections