Run a 5 ps simulation on 1EBY_dimer_ligand_BEB.pdb, poll every
7 minutes and visualise the results.
Everything else is clean: 2 chains, 99 residues each, no altlocs, no insertion codes, no missing-residue gaps. The 95 crystallographic waters get replaced with fresh TIP3P, which is expected.
The fix: delete the HETATM records and supply BEB as a separate posed 3D SDF. I can do that here — it needs bond-order perception, so I would install RDKit first.
Do that, then try again.
Bond orders come from the PDB Chemical Component Dictionary rather than a guess. All 48 atom names matched the CCD entry exactly.
| Extracted | 48 heavy atoms, 93 bonds |
| Hydrogens | 40 added, MMFF-relaxed with heavy atoms fixed |
| Formula | C38H40N2O8 — matches CCD |
The docked pose is untouched — only hydrogen positions were relaxed.
Ready to submit — confirm before I start.
| Setting | Value |
|---|---|
| Input | 1EBY_receptor.pdb + BEB.sdf |
| Simulation time | 5 ps — 2500 steps @ 0.002 ps |
| Ensemble | NVT — minimize → equilibrate → produce |
| Force field | ff19SB / TIP3P (default) |
| Platform | CUDA (default) |
| Estimated | ~4 min · ~4 credits |
Submit.
| Job ID | 21173bb3-c386-4a31-b0d0-d7476cc469c6 |
| Params | simulation_time_ps: 5 |
| Status | running |
Success. 5 ps NVT, ff19SB/TIP3P, OpenMM 8.0 on CUDA.
| Metric | Value |
|---|---|
| Temperature | 300.73 ± 1.74 K |
| Cα RMSD | 0.59 Å final · 0.62 max |
| Mean Cα RMSF | 0.35 Å |
| PE drift | 0.20 % |
| System | 198 res · 3134 protein + 88 ligand atoms |
Energy bounded, temperature held, structure intact. The Asp25/Asp25′ catalytic dyad holds the inhibitor diol at ~100 % occupancy across all 125 frames.
Viewer is on the right — switch between the ligand pocket and the chain interface to see contacts with live distances.
At 5 ps this shows the system is stable and correctly built. It is far too short for binding conclusions.
Example session, lightly condensed. The viewer alongside is the real output of job 21173bb3 — 125 frames of a 5 ps run, fully interactive and running entirely in this page.