Quickstart: mlmm all

Goal

Run the end-to-end ML/MM ONIOM workflow once from a reactant and product PDB pair: MM parametrization → ML-region selection → MEP search (single-pass path-opt by default) → optional TS optimization, thermochemical correction, IRC, and DFT single-point.

Prerequisites

  • Two complex PDBs with matching atom order and identical residues (R/P; only coordinates differ — in PyMOL tick Original atom order on export).

  • Explicit hydrogens present (mlmm all does not auto-protonate); match -l RES:CHARGE to the H count in the file (e.g. SAM 23 H = SAM:1, 22 H = SAM:0; a mismatch triggers an antechamber sqm odd-electron failure).

  • GPU recommended. Default backend uma (also orb / mace / aimnet2 via -b).

Minimal command

End-to-end MEP only (matches examples/toy_system/run.sh test18):

mlmm all -i r_complex.pdb p_complex.pdb -c PRE -r 6.0 \
  --ligand-charge 'PRE:0' -q -1 -m 1 --out-dir ./result_all

With TS optimization, thermochemistry, and DFT single-point:

mlmm all -i r_complex.pdb p_complex.pdb -c PRE -r 6.0 \
  --ligand-charge 'PRE:0' -q -1 -m 1 \
  --tsopt --thermo --dft --out-dir ./result_all

(-c = ligand residue list for ML-region centring; -r = ML-region radius in Å; -q/-m = total system charge / multiplicity.)

Result

Read summary.log (human-readable) and summary.json (segments, ΔE‡, ΔE, bond_changes); the merged path is mep.pdb / energy_diagram_MEP.png at the output root. See all for the full output tree and per-segment deliverables, and output-layout for the filename reference.

Next step