mk_prepare_receptor: keep protonation of the input structure

[DrrDom]

If a protein structure was prepared elsewhere and has all hydrogens explicitly defining specific protonation states, will meeko keep them as is or it may alter them according to some internal logic during the conversion to pdbqt format?

I'm curious whether we can safely convert previously prepared structures into pdbqt format using meeko trusting that all input protonation states will be kept untouched?

[diogomart]

Yes, meeko keeps the input Hs, unless option --set_template is used to explicitly define protonation state, e.g., --set_template A:66=HID

Non-polar Hs are merged by default.

[DrrDom]

Thank you!

One more note on protonation states. If a C-terminal residues is protonated it returns an error. If I remove the hydrogen on COOH group of the last residue it works fine. I do not know whether it is a bug or a feature.

mk_prepare_receptor.py -i 5.pdb -o 5 -p --wanted_altloc A:118=A,A:177=B,A:232=A

@> 4534 atoms and 1 coordinate set(s) were parsed in 0.02s.
No template matched for residue_key='A:298'
tried 3 templates for residue_key='A:298'excess_H_ok=False
LEU        heavy_miss=0 heavy_excess=1 H_excess=[-1] bond_miss={0} bond_excess=set()
NLEU       heavy_miss=0 heavy_excess=1 H_excess=[-1] bond_miss={0} bond_excess={4}
CLEU       heavy_miss=0 heavy_excess=0 H_excess=[2] bond_miss=set() bond_excess=set()

Error: Creation of data structure for receptor failed.

Details:
- Template matching failed for: ['A:298']
These residues can be ignored with option allow_bad_res.

Recommendations:
1. (for batch processing) Use -a/--allow_bad_res to automatically remove residues
that do not match templates, and --default_altloc to set
a default altloc variant. Use these at your own risk.

2. (processing individual structure) Inspecting and fixing the input structure is recommended.
Use --wanted_altloc to set variants for specific residues.

5.pdb.txt

[diogomart]

Meeko doesn't have residue templates for C-terminal carboxylic acids (COOH), only for carboxylates. I think it also doesn't have templates for neutral amine N-terminus, only for protonated ones. There's some capping groups that Charmm-GUI supports that Meeko doesn't. Meeko supports only ACE (acyl) and NME (methylamine), for N- and C-terminus, respectively. We should add these templates.

[diogomart]

Here are the C-terminus variants that Meeko currently supports:

Negatively charged. This includes the OXT oxygen (Amber naming)

C   (alpha)
 \
  C = O
  |
  O-   (OXT)

Blunt ending. Neutral, chemically unreasonable as an H is missing, but the charges sum up to an integer.

C   (alpha)
  \
   C = O

NME capped. NME is a separate residue. (Charmm-GUI puts everything in the same residue, I think, and Meeko currently can't use those models).

C   (alpha)
  \
   C = O
   |
 H-N
    \
     CH3

[DrrDom]

Thank you! Now it's clear for me.