Output Structure

QuantumPDB organizes output into a directory hierarchy under the path specified by output_dir. Each PDB gets its own subdirectory containing all pipeline outputs.

Directory Layout

After running all three stages (qp run, qp submit, qp analyze), a typical output directory looks like:

output_dir/
└── {pdb}/
    ├── {pdb}.pdb                    # Original fetched PDB
    ├── {pdb}.ali                    # Modeller alignment file
    ├── {pdb}_modeller.pdb           # Rebuilt structure (Stage 1)
    ├── Protoss/
    │   ├── {pdb}_protoss.pdb        # Protonated structure (Stage 2)
    │   ├── {pdb}_protoss_orig.pdb   # Copy before active-site fixes
    │   ├── {pdb}_ligands.sdf        # Ligand structures (SDF format)
    │   └── {pdb}_log.txt            # Protoss processing log
    ├── charge.csv                   # Per-residue charges
    ├── count.csv                    # Residue counts per sphere
    ├── spin.csv                     # Radical species spins (if present)
    └── {CENTER}_{CHAIN}{ID}/        # One directory per center residue
        ├── 1/                       # Sphere 1 (innermost)
        │   ├── cluster.pdb
        │   └── cluster.xyz
        ├── 2/                       # Sphere 2 (includes sphere 1)
        │   ├── cluster.pdb
        │   └── cluster.xyz
        └── ...                      # Additional spheres

Cluster Directory Naming

Cluster directories are named after the center residue using the format {RESNAME}_{CHAIN}{RESID}. For example:

  • FE_A501 — iron at position 501 on chain A

  • CU_A357-CU_A358 — merged dicopper center

  • HEM_A155 — heme group

Each numbered subdirectory contains a progressively larger cluster model. Sphere 1 is the innermost (first coordination shell), sphere 2 adds the next layer of interacting residues, and so on. Higher-numbered spheres are supersets — sphere 2 contains all atoms from sphere 1 plus the second shell.

File Descriptions

Structure files:

  • {pdb}.pdb — The original PDB file, either downloaded from the RCSB or copied from a local path.

  • {pdb}_modeller.pdb — The structure after Modeller has rebuilt missing atoms, residues, and loops.

  • {pdb}_protoss.pdb — The protonated structure with hydrogens added, alternate conformations resolved, and active-site corrections applied.

  • {pdb}_protoss_orig.pdb — A copy of the Protoss output before metalloenzyme-specific active-site corrections.

Ligand data:

  • {pdb}_ligands.sdf — SDF file containing ligand structures as output by Protoss. Used to compute ligand charges and spins.

Charge and count data:

  • charge.csv — Net charge for each residue in the cluster. Columns: chain ID, residue name, residue ID, net charge. Includes both amino acid charges (from protonation state analysis) and ligand charges (from Protoss SDF).

  • count.csv — Number of residues in each interaction sphere.

  • spin.csv — Spin contributions from radical species (e.g., NO as doublet, O2 as triplet). Only generated when radical ligands are present.

Cluster models:

  • cluster.pdb — QM cluster model in PDB format with chain break caps.

  • cluster.xyz — QM cluster model in XYZ format (suitable for direct use with QM codes).

QM job files (after qp submit):

  • QM input scripts (TeraChem .in or ORCA .inp)

  • Scheduler submission scripts (.sh)

  • .submit_record — Hidden marker file created after job submission to prevent duplicate submissions

  • ptchrges.xyz — MM point charges for charge embedding (if enabled)

Analysis output (after qp analyze):

  • checkup/failure_modes.csv — Failure mode classification for all jobs

  • checkup/job_status.csv — Status of each job

  • checkup/failure_modes.png — Plot of failure modes

  • checkup/job_status.png — Plot of job statuses