pdb2mcsm

English    한국어

Clean and prepare protein complex PDB files for mCSM-PPI2 ΔΔG analysis.

MD simulation outputs and other "raw" structure files often can't be fed directly into mCSM-PPI2: chains may be merged or unlabeled, residues renumbered from 1, waters and ions cluttering the file, and histidines written in MD-specific names like HIE. pdb2mcsm fixes all of this automatically so the structure is accepted by the server.

Try it in Colab (no install)

Upload a PDB, get a cleaned _ready.pdb back, and optionally generate an alanine-scanning mutation list — all through form fields, no coding required.

What it does

Automatic cleanup

• Chain splitting by priority: existing chain IDs → TER records → residue-number jumps (flagged when guessed).
• Handles any number of chains (dimer, trimer, antibody–antigen, ...).
• Removes water, ions, and common crystallization additives (HOH, NA, CL, SO4, GOL, ...).
• Strips hydrogens (mCSM re-adds them).
• Normalizes MD residue names: HIE/HID/HIP → HIS, CYX → CYS, MSE → MET, and more.
• Handles phosphorylated residues: SEP/S1P → SER, TPO/T1P → THR, PTR/Y1P → TYR, with phosphate atoms removed.
• Preserves insertion codes (e.g. 100A).
• Assigns clean chain IDs (keeps originals when valid).

Optional

• Renumber a chain by a fixed offset (e.g. shift file numbering to paper numbering).
• Generate an alanine-scanning mutation list in mCSM-PPI2 batch format.

Safety — the tool warns instead of silently guessing: ambiguous chain splits and single-chain inputs raise clear warnings so you can verify before trusting the output.

Quick start (Python)

import core

# 1. Read and split chains
header, chains, warnings, method = core.parse_pdb("my_complex.pdb")
core.assign_chain_ids(chains)

# 2. Review (important!)
print("split method:", method)
for row in core.summarize(chains):
    print(row)          # (chain, #residues, first, last, split_reason)
for w in warnings:
    print("WARNING:", w)

# 3. (optional) shift a chain's numbering: file 36 -> paper 98 means offset 62
core.renumber_chain(chains, "B", 62)

# 4. Write the cleaned file -> upload this to mCSM-PPI2
core.write_clean("my_complex_ready.pdb", header, chains)

# 5. (optional) alanine-scanning list for the batch box
print("\n".join(core.build_ala_scan(chains, "B", 95, 110)))

Then in mCSM-PPI2

1. Open the mCSM-PPI2 batch/multiple-mutation page.
2. Upload my_complex_ready.pdb.
3. Select the chain to mutate.
4. Paste the alanine-scanning list.
5. Run → per-residue ΔΔG. Large negative values mark residues important for binding.

Functions

Function	Purpose
parse_pdb(path)	Read a PDB; split into chains; return (header, chains, warnings, method)
assign_chain_ids(chains)	Give each chain a clean A/B/C ID
write_clean(path, header, chains)	Write the cleaned PDB (H stripped, phosphate atoms removed)
renumber_chain(chains, chain_id, offset)	Shift a chain's residue numbers by offset
build_ala_scan(chains, chain_id, start, end)	Build an mCSM-PPI2 alanine-scanning list
summarize(chains)	Per-chain summary rows for review
read_atoms(path)	Low-level: parse ATOM/HETATM records into Atom objects

Limitations

pdb2mcsm solves file structure and cleanup problems. It does not handle:

• Automatic renumbering — you must know the offset yourself (renumber_chain is a fixed shift).
• Other modified residues beyond phosphorylation (acetylation, methylation, glycosylation, ...) — add them to NONSTD_MAP as needed.
• Ligands / cofactors (drugs, heme, ATP, metal clusters) — only water/ions are removed.
• Structural defects — missing atoms, broken backbone, atom clashes, or alternate locations (altloc) are not repaired.
• Nucleic acids — protein–DNA/RNA complexes are out of scope (mCSM-PPI2 is protein–protein only).

Some limits are mCSM-PPI2's own, not this tool's: phosphorylation effects are lost once a residue is converted to its standard form; protein–ligand and protein–nucleic-acid binding aren't supported; very large complexes may exceed server limits.

Notes

• Default keeps original residue numbering — the numbers you feed mCSM must match what's in the cleaned file. Use summarize() or inspect chains[i].residues to confirm.
• Converting a phosphorylated residue to its standard form removes the phosphorylation from the model. If the phosphorylation itself matters for your binding question, interpret results accordingly or use a method that supports modified residues.

License

MIT