STRC h01 Phase 9b-lite — Residue-Field Decomposition

Phase 9b-lite supports the broad electrostatic sign: E1659A-WT field delta +0.181 q/A at 18 A, p=1.5e-10; 9 A inner shell opposes.

Problem

Phase 9b proper is holo MD plus MM-PBSA, but gmx_MMPBSA is not installed in the local PATH yet. This smoke proof asks the tractable post-processing question first: using existing matched WT and E1659A MD snapshots, does the formal side-chain charge environment around the K1141 pocket independently preserve the E1659A > WT electrostatic sign without APBS grids, Vina scores, Boltz-2, or tauRAMD?

Method

  • Script: hypotheses/h01-pharmacochaperone/scripts/phase9b_lite_residue_field_decomposition.py
  • Inputs:
    • WT ensemble: models/artifacts/phase5a_snapshots/snap_000-019.pdb
    • E1659A ensemble: models/artifacts/phase5d_snapshots/snap_000-019.pdb
  • Pocket center: centroid of K1141-ring C-alpha atoms from the existing h01 pocket definition: real residues 1141, 1135, 1137, 1165, 1167, 1175. WT ultra-mini residue numbers use the existing Phase 5a offset of -593.
  • Charge proxy: one formal side-chain representative per charged residue: Lys NZ +1, Arg CZ +1, Asp CG -1, Glu CD -1. Histidine is omitted rather than assigning an ambiguous protonation state.
  • Readout: sum q/r in e/A for charge representatives within sensitivity radii 9, 12, 15, 18, 21, and 24 A of the pocket center. Matched snapshots are compared by paired t-test.

Parameter provenance

ParameterValueSource / rationale
Snapshot countn=20 WT + n=20 E1659AExisting Phase 5a / Phase 5d h01 MD ensembles.
Pocket ring residues1141, 1135, 1137, 1165, 1167, 1175Existing h01 K1141 pocket-ring definition used in Phase 5e / Phase 5k analyses.
WT residue offset-593Existing Phase 5a ultra-mini construct numbering convention.
Radii9-24 A sensitivity gridMethod sensitivity sweep, not a biological constant.
Formal side-chain chargesLys/Arg +1; Asp/Glu -1Standard formal-charge decomposition proxy; used only for sign / residue-field smoke, not physical free energy.

Results

Radius (A)WT field q/AE1659A field q/ADelta MUT-WT q/Apaired t pWT net qMUT net q
9-0.0271 +/- 0.0018-0.1120 +/- 0.0059-0.0849 +/- 0.00633.80e-110.00-0.95
12-0.1214 +/- 0.00200.0372 +/- 0.0104+0.1587 +/- 0.01099.61e-12-1.000.60
15-0.0731 +/- 0.01330.0079 +/- 0.0060+0.0810 +/- 0.01575.53e-05-0.300.15
18-0.0677 +/- 0.00740.1133 +/- 0.0125+0.1810 +/- 0.01461.54e-10-0.201.85
21-0.1721 +/- 0.01480.1645 +/- 0.0102+0.3366 +/- 0.01975.44e-13-2.303.00
24-0.3495 +/- 0.00760.1328 +/- 0.0102+0.4823 +/- 0.01411.56e-18-6.202.25

Artifacts:

  • hypotheses/h01-pharmacochaperone/artifacts/phase9b_lite_residue_field/phase9b_lite_residue_field.json
  • hypotheses/h01-pharmacochaperone/artifacts/phase9b_lite_residue_field/radius_summary.csv
  • hypotheses/h01-pharmacochaperone/artifacts/phase9b_lite_residue_field/per_snapshot_radius_fields.csv
  • hypotheses/h01-pharmacochaperone/artifacts/phase9b_lite_residue_field/phase9b_lite_residue_field.md

Verdict

  • Outcome: SMOKE_PASS for the broad charged-neighbourhood sign.
  • Key number: 18 A shell delta = +0.1810 +/- 0.0146 q/A, paired p=1.54e-10.
  • Interpretation: A force-field-light, APBS-independent decomposition preserves the main Phase 5k claim that the E1659A pocket neighbourhood is more favorable for an anion-like ligand than WT.
  • Limitation: The 9 A inner shell opposes the sign. The mechanism is not a uniformly positive microenvironment; the supportive signal is a wider charged-neighbourhood effect from 12 A outward. This proof is not a physical binding free energy and does not replace 9b production MM-PBSA.

Ranking delta

  • Hypothesis h01: tier A A | mech 4 4 | deliv 3 3 | misha_fit 4 4.
  • No hub frontmatter change. This is supportive orthogonal smoke evidence, not a production free-energy track.
  • Next compute: keep Phase 9b production pending until gmx_MMPBSA or an AmberTools MM-PBSA path is installed; meanwhile this result counts as a low-cost residue-decomposition cross-check, not one of the 5 production Phase 9 track verdicts.

Connections