Misidentification of Ligand Atoms vs Protonation

[CLG68]

see the second message below, this seems to be a problem with deprotonation...

---

Please note that the description below was generated by Gemini 2.5 PRO that I'm using to help me implementing FreeSASA analysis of buried SASA of the interacting atoms. I uploaded an archive with the receptor complex and the report I get from the PLIP server.

Context

I've encountered a situation with PLIP (both the web server v2.4.0 and a local library installation) where it appears to be reporting PDB atom serial numbers belonging to protein residues in fields that are expected to denote ligand atoms for certain interaction types. This occurs even when the ligand itself is correctly identified by its residue name, chain, and number.

for https://plip-tool.biotec.tu-dresden.de/plip-web/plip/index, I used latest version of Firefox under Ubuntu 24.04.2 LTS

Context:

PDB File: A protein-ligand complex that has undergone energy minimization (Z_complex-minimized.pdb).
Ligand: Identified as MOL C 295 (HETATM records for this ligand start at atom serial 4657 and it has 68 atoms).
PLIP Version: Web server (v2.4.0 used on 2025/05/11) and a local Python library installation show consistent behavior.

Observed Issue with Z_complex-minimized.pdb:

Hydrogen Bonds:
    PLIP correctly identifies the ligand as MOL C 295.
    For an H-bond interaction involving protein residue ASP A 87, the PLIP report (and my script's parsing of the local PLIP library output) indicates PROTISDON: False (meaning the ligand is the donor).
    The DONORIDX (expected to be the ligand's donor atom) is reported as PDB atom serial 2299.
    Problem: In my PDB file, atom serial 2299 is N VAL A 140, clearly a protein atom, not part of ligand MOL C 295. The ligand MOL C 295 only has 68 atoms, with serials starting much higher (e.g., 4657+).

Pi-Stacking Interactions:
    For a pi-stacking interaction involving protein residue PHE A 245 and ligand MOL C 295.
    The LIG_IDX_LIST (expected to contain PDB atom serials of the ligand's aromatic ring) is reported as [2305, 2306, 2307, 2316, 2317, 2318].
    Problem: These serial numbers also correspond to protein atoms (e.g., atom 2305 is H VAL A 140), not atoms from the ligand MOL C 295. The PROT_IDX_LIST correctly identifies atoms from PHE A 245.

Impact:
This misidentification of "ligand" atoms (by providing protein atom serials in ligand-designated fields) makes it impossible to perform accurate downstream analysis that relies on knowing the specific ligand atoms involved in an interaction. For example, calculating the Solvent Accessible Surface Area (SASA) changes for only the ligand atoms participating in a specific interaction becomes erroneous if the provided serials point to protein atoms. My script, which attempts this, correctly identifies that these serials are not part of the isolated ligand, leading to a calculated free SASA of 0.0 for these specific interaction components.

Expected Behavior:
For a given interaction with a defined ligand (e.g., MOL C 295), attributes like DONORIDX (when PROTISDON is false) or LIG_IDX_LIST should contain PDB atom serial numbers that actually belong to that specified ligand (MOL C 295).

Thank you for your help

PLIP-report_and _complex.zip

[CLG68]

I tried the corrected file from Bug in result #167 and it worked. So I transformed my pdb to look like it and then I used the PLIP server. What it did..."fixed" the pdb by deprotonating protein, renumbering the atoms w/o updating the connection table. It assigned a chain ID: Z to the ligand. This renumbering explain why the ligand atom number did not fit with anything.

The file from Bug in result #167 was not deprotonated. Why is my file deprotonated and not this one? Why is the PLIP server accepting a deprotonated complex as input for evaluating interactions?

my modified complex-files:

modified-files.zip

Update:

I listed all the the arguments using: python3 -m plip.plipcmd -h

Then I use : --nohydro & --nofix

python3 -m plip.plipcmd -f Z_complex-minimized.pdb --nohydro --nofix -x -o plip_report_no_hydro_no_fix.xml

This prevented the deprotonation and renaming of my file that was, anyway, already processed/protonated by pdb2pqr and minimized.

I do not think those are option available on the PLIP server. Minimally, the server should verify that the fixed pdb is protonated before running a profile of interaction or at least clearly inform the user the profile was done on a non-protonated complex.