STRC h01 Phase 9b — MM-PBSA Tooling Preflight

gmx_MMPBSA / AmberTools are installed and callable, but Phase 9b differential MM-PBSA is blocked by missing matched inputs, not by missing software.

Question

Can we run the planned Phase 9b WT-vs-E1659A MM-PBSA / MM-GBSA smoke directly from existing h01 assets?

Script: hypotheses/h01-pharmacochaperone/scripts/phase9b_mmpbsa_smoke.py
Tooling envs:
- AmberTools / gmx_MMPBSA: /opt/miniconda3/envs/strc-mmpbsa
- OpenMM / ParmEd loader: /opt/miniconda3/envs/strc-mmgbsa
Checked executables: gmx, gmx_MMPBSA, MMPBSA.py, cpptraj, tleap, antechamber.
Attempted input path: Phase 5q E1659A holo OpenMM system for v5.3__aq3__adamantyl__acylsulfonamide_SO2Me__-Cl.

Tooling is ready:

Tool	Status
GROMACS	`/opt/homebrew/bin/gmx`, 2026.1-Homebrew
gmx_MMPBSA	`/opt/miniconda3/envs/strc-mmpbsa/bin/gmx_MMPBSA`, 1.6.4
MMPBSA.py	callable with `AMBERHOME=/opt/miniconda3/envs/strc-mmpbsa`
cpptraj	AmberTools V6.18.1
tleap / antechamber	callable

Input blockers:

Gate	Result
Native h01 GROMACS / Amber files	none found (`.top`, `.tpr`, `.xtc`, `.trr`, `.gro`, `.prmtop`, `.inpcrd`)
E1659A holo v5.3 OpenMM system	found: `phase5q_strc_v53/per_ligand/.../system.xml`, `topology.pdb`, DCD replicas
Matched WT holo v5.3 system	not found
OpenMM → Amber conversion	blocked: ParmEd loads 56,785 atoms / 18,337 residues, but cannot write valid Amber topology because 37,405 bonds lack Amber bond types

Artifacts:

hypotheses/h01-pharmacochaperone/artifacts/phase9b_mmpbsa_smoke/phase9b_mmpbsa_smoke.json
hypotheses/h01-pharmacochaperone/artifacts/phase9b_mmpbsa_smoke/phase9b_mmpbsa_smoke.md

Outcome: TOOLING_READY_INPUTS_BLOCKED.
Meaning: the Phase 9b smoke is now blocked by matched-input construction, not by local installation.
Do not count this as a Phase 9 production track verdict. It produced no WT-vs-E1659A MM-PBSA sign.
Next P0: build matched WT and E1659A holo systems for the same v5.3 lead through a single topology path, preferably AmberTools / OpenFF from the start rather than post-hoc conversion of OpenMM XML.

Hypothesis h01: tier A → A | mech 4 → 4 | deliv 3 → 3 | misha_fit 4 → 4.
No score move. This is a tooling/input preflight and blocker note, not scientific support for or against the pharmacochaperone differential.
Hub next_step should now prioritize matched WT/MUT holo topology generation before any MM-PBSA production claim.