Phase 8h-lite · Light computational evidence package for H01 (paper-ready add-ons)
Context. H01 mech-4 is paper-ready (Phase 5k matched ensemble APBS, +5.99 vs +1.46 kT/e MUT vs WT, p=6.9e-12, Cohen’s d=3.10). Lead v5.2__aq3__adamantyl__CONHOH__-Cl is ADMET-clean. Heavy computational follow-up = Boltz-2 28-job batch (running). Light add-ons here close five paper-claim gaps using existing data + RAG-mined literature, no new MD or docking.
Done
| # | Question | Method | Verdict | Closes |
|---|---|---|---|---|
| 1 | Is K1141 pocket druggable? | Halgren SiteMap-style Python grid burial estimator | YES — V 1145 ų, phobic 0.61, philic 0.39, V_pocket/V_lig 2.75; aromatic-rich shell (W1612, F1646, F1169, W1652) | Phase 5c GREEN with quantitative number |
| 2 | Can the lead cross the round-window membrane as ear-drop? | Salt 2001 baseline + Stokes-Einstein MW + Avdeef logP/TPSA + Henderson-Hasselbalch | YES — predicted P 7.6× TMPA → ~60% applied conc in basal turn at 90 min, robust to pKa ±0.5 | Delivery claim “topical RWM ear-drop” |
| 3 | Selectivity vs cochlear off-target K/R pockets | Charge proxy at Vina pose + receptor-wide K/R cluster scan | INCONCLUSIVE — proxies don’t capture enclosure, off-targets have many K/R clusters but in surface-exposed regions; needs APBS on enclosed pockets (heavy planned) | Negative — flag for proper APBS step |
| 4 | Is Vina lead pose ensemble-stable against Phase 5d K1141 motion? | Distance K1141 NZ ↔ lead CONHO⁻ over 20 snapshots | YES — K1141 NZ SD 1.1 Å, lead CONHO⁻↔NZ 5.02 ± 0.57 Å (in Coulomb attraction range); APBS-correct binding mode would tighten to ~3 Å salt-bridge | Mech 4 holds for the lead’s actual geometry |
| 5 | Lead similarity to known cochlear ototoxins? | Morgan FP r=2 + Tanimoto vs 12-compound class panel from Schacht 2008 | NO OVERLAP — max 0.127 (aspirin), threshold 0.40; no aminoglycoside/loop-diuretic/platinum/macrolide motifs | Chemical-class novelty for lead |
Compute footprint
| step | time | cores | output |
|---|---|---|---|
| #1 druggability grid | 15 s | 1 | result_01_druggability.md |
| #2 RWM permeability | <1 s | 1 | result_02_rwm_permeability.md |
| #3 selectivity proxy | 3 s | 1 | result_03_selectivity_scan.md |
| #4 pose stability | <1 s | 1 | result_04_pose_stability.md |
| #5 Tanimoto | 1 s | 1 | result_05_tanimoto.md |
| total | <30 s |
What it adds to the paper
New paper-claim cells unlocked:
- Quantitative druggability of K1141 pocket (table value, not adjective).
- Predicted topical-delivery PK with sensitivity envelope.
- Ensemble-validated mech-4 geometry on the actual lead (not generic anion probe).
- Chemical-class clean tox baseline.
Still required (heavy planned):
- Boltz-2 28-job batch for receptor-flexibility-aware re-rank (running).
- APBS on each off-target enclosed pocket for proper selectivity claim.
- Wet-lab cytotoxicity panel (HEI-OC1, UB-OC1).
Connections
[part-of]h01-pharmacochaperone[builds-on]STRC h01 Phase 5k-WT Matched Ensemble APBS 2026-04-25[builds-on]STRC h01 Phase 8d 8e 8f v5.2 Library Design 2026-04-25[blocked-by]STRC Computational Scripts Inventory — to be appended