Quickstart: pdb2reaction all (Endpoint mode)

Goal

Run the end-to-end workflow once from two full PDB structures.

Prerequisites

  • pdb2reaction installed (see Installation)

  • Two PDB files (reactant R and product P) with hydrogen atoms already added

  • The same atoms in the same order across all input PDB files

About the example filenames: 1.R.pdb and 3.P.pdb mirror the numbered reactant/product files shipped in the geranyl pyrophosphate (GPP) C6-methyltransferase BezA example directory (examples/ — 1.R.pdb = reactant state, 3.P.pdb = product state, with intermediate 2.*.pdb files for multi-step runs). Replace them with the two (or more) full-system PDBs for your own reaction. To run the commands below verbatim, first fetch the bundled example: git clone https://github.com/t-0hmura/pdb2reaction && cd pdb2reaction/examples.

Minimal command

pdb2reaction all -i 1.R.pdb 3.P.pdb -c 'SAM,GPP,MG' -l 'SAM:1,GPP:-3' \
 --out-dir ./result_all

(Optional) Add post-processing in the same run

pdb2reaction all -i 1.R.pdb 3.P.pdb -c 'SAM,GPP,MG' -l 'SAM:1,GPP:-3' \
 --tsopt --thermo --dft --out-dir ./result_all

VRAM warning: --dft launches GPU4PySCF single-point jobs on the extracted cluster model and can easily OOM on GPUs with < 24 GB VRAM for clusters above ~200 atoms. If you hit CUDA out of memory, either drop --dft and run pdb2reaction dft separately with a smaller basis / trimmed cluster, or move the DFT step to a larger-VRAM node. The [dft] extra must also be installed (see Installation Step 7).

Expected output

A successful run produces a directory like:

result_all/
├── summary.log                    # Human-readable summary
├── summary.json                   # Machine-readable results
├── mep.pdb                        # Merged MEP path (promoted to the root)
├── energy_diagram_MEP.png         # All-segment MEP energy profile
├── segments/
│   └── seg_01/                    # Per-reactive-segment deliverables
│       ├── reactant.pdb           # Canonical IRC-optimized R/TS/P
│       ├── ts.pdb
│       ├── product.pdb
│       ├── ts/final_geometry.pdb  # Present with --tsopt
│       ├── irc/finished_irc_trj.xyz
│       └── freq/                  # Present with --thermo
└── _work/                         # Pipeline scratch (safe to delete)
    └── path_opt/                  # Raw MEP-engine output (path_search/ with --refine-path True)
        └── summary.json           # MEP engine results

What to check:

  1. summary.json — check the status field ("success", "partial", or "failed") and the per-segment barrier_kcal values; summary.log presents the same information in human-readable form

  2. segments/seg_01/*.pdb — open in PyMOL to verify the R/TS/P structures make chemical sense

  3. energy_diagram_*.png — the energy profile should show a clear barrier

Sample terminal output (successful run):

[all] Elapsed for Whole Pipeline: HH:MM:SS.sss

(Wall-clock varies with system size, GPU, and selected stages.)

If --tsopt is enabled, you should also see:

[Imaginary modes] n=1  ([-425.9])

A first-order saddle point shows exactly one imaginary mode along the reaction coordinate. IRC validation (run automatically as part of --tsopt) confirms it connects the expected reactant and product.

Tips

  • pdb2reaction all --help shows core options; pdb2reaction all --help-advanced shows the full list.

Next step