STRC Research

STRC h01 Phase 9b Matched Amber Input Compatibility 2026-04-27

2 min read

STRC h01 Phase 9b — Matched Amber Input Compatibility

Matched WT/E1659A dry Amber inputs now build and MMPBSA.py reads one frame for both, but huge VDWAALS terms make the one-frame energies nonphysical.

Question

After the Phase 9b preflight showed that endpoint-free-energy tooling was installed but matched inputs were missing, can h01 now produce matched WT and E1659A Amber dry complexes that the endpoint tool actually reads?

Method

  • Builder: hypotheses/h01-pharmacochaperone/scripts/phase9b_matched_holo_input_builder.py
  • Artifact root: hypotheses/h01-pharmacochaperone/artifacts/phase9b_matched_holo_inputs/
  • Ligand: v5.3__aq3__adamantyl__acylsulfonamide_SO2Me__-Cl
  • Snapshot: snap_010
  • Parameter path: AmberTools / GAFF2 via antechamber, parmchk2, and tleap in /opt/miniconda3/envs/strc-mmpbsa.
  • Endpoint smoke: one-frame MMPBSA.py GB run (igb=5, saltcon=0.150) on WT and E1659A dry complex/receptor/ligand prmtops.

Results

CheckWTE1659A
Builder statusPASSPASS
Residue sanity at 1141LYSLYS
Residue sanity at 1659GLUALA
K1141-ring alignment RMSD0.877 A0.000 A
tleapErrors 0 / Warnings 4Errors 0 / Warnings 4
MMPBSA.py frames processed11
MMPBSA.py wall time2.120 min1.863 min
DELTA TOTAL1.7052441820e10 kcal/mol1.3075101011e8 kcal/mol

Core outputs:

  • artifacts/phase9b_matched_holo_inputs/wt/{complex,receptor,ligand}.{prmtop,inpcrd}
  • artifacts/phase9b_matched_holo_inputs/e1659a/{complex,receptor,ligand}.{prmtop,inpcrd}
  • artifacts/phase9b_matched_holo_inputs/wt/FINAL_RESULTS_MMPBSA_1FRAME.dat
  • artifacts/phase9b_matched_holo_inputs/e1659a/FINAL_RESULTS_MMPBSA_1FRAME.dat

Verdict

  • Outcome: INPUTS_READY_ENDPOINT_READ_PASS_NUMBERS_INVALID.
  • Resolved layer: Phase 9b is no longer blocked at matched dry Amber input construction.
  • Endpoint compatibility: MMPBSA.py reads both WT and E1659A prmtop/inpcrd/NetCDF sets and finishes normally.
  • Scientific limitation: the one-frame endpoint numbers are not interpretable. The gigantic positive VDWAALS terms mean the smoke found a geometry/relaxation problem, not a binding free-energy sign.
  • Next layer: minimize/equilibrate the matched holo complexes, inspect ligand/protein clashes and intended ligand charge state, then rerun the Phase 9b smoke on relaxed frames before any WT-vs-E1659A claim.

Ranking delta

  • Hypothesis h01: tier A → A | mech 4 → 4 | deliv 3 → 3 | misha_fit 4 → 4.
  • No score move. This closes the input/tool-readiness blocker only; it does not produce a valid MM-PBSA differential.
  • Hub next_step should move from “build matched inputs” to “relax matched inputs and rerun endpoint smoke on sane frames.”

Connections

Graph View

Backlinks