All 50 docks (10 conditions x 5 seeds) used identical box / scoring
parameters. The literal CLI for one example — R215A_N226A_holo, seed 42 — is:
/opt/homebrew/bin/vina \
--receptor /Users/ario/conserved_site_project/07e_mut_docking_v5/R215A_N226A/R215A_N226A_holo.pdbqt \
--ligand /Users/ario/conserved_site_project/05b_ligand_v2/dump.pdbqt \
--center_x -0.137 --center_y 4.232 --center_z 15.159 \
--size_x 18.0 --size_y 18.0 --size_z 18.0 \
--exhaustiveness 32 \
--num_modes 20 \
--seed 42 \
--out /Users/ario/conserved_site_project/12_phase7/01_replicas/raw/R215A_N226A_holo_seed42.pdbqt
| Parameter | Value |
|---|---|
| Centroid x | -0.137 |
| Centroid y | 4.232 |
| Centroid z | 15.159 |
| Box size x/y/z | 18.0 A x 18.0 A x 18.0 A |
| Exhaustiveness | 32 |
| num_modes | 20 |
| Seeds | 42, 7, 13, 99, 256 |
| CPU | 4 (auto) |
| Receptor format | PDB2PQR-corrected -> Meeko PDBQT |
| Ligand | dUMP, 5 active torsions (heavy=20) |
The centroid is the chain-A active-site Cα centroid carried over from
the v3/v4/v5 canonical pipeline (the v5 mutant runs that produced
07e_mut_docking_v5/mutant_results_v5.csv). The 18-A cubic box is
slightly smaller than the v5 22-A canonical box; this is intentional —
Phase 7 is specifically asking for stochastic noise floor of a tight
box around the dUMP pocket, not a wide rescore.
source 00_setup/env.sh.--out so the docked poses are
captured. The --seed flag fully determines the run, so identical
commands are bit-reproducible on the same Vina build.