[gmx-users] Defining protonation in pdb2gmx input

[mhuhtala at abo.fi]

This is continuing a thread from a month ago.

> Quoting Justin Lemkul <jalemkul at vt.edu>:

>> Now I did a bit more testing, and at least renaming ASP to ASPH in  
>> the input will get me an ASPH in the output, but only the residue  
>> name, not the hydrogen. So I get a residue named ASPH, but the  
>> structure is an unprotonated ASP. Is this a bug? Or is pdb2gmx  
>> expecting to find the hydrogen atom line in the input?


> All necessary hydrogens should be reconstructed according to the  
> .hdb file (for which entries for protonated amino acids exist and  
> are functional).

Reading the source code of pdb2gmx, that does not seem true. I don't  
fully understand the code after a quick look, but it seems to me that  
ASPH and GLUH are always matched to ASP and GLU (if not entered  
manually), and that the protonation routine is only for HIS. And once  
pdb2gmx matches ASPH with ASP, it refuses to read the hydrogen atom  
even if it is in the input (because it wants ASP).

>> This is all using Gromacs 2018 and specifying GROMOS96 54a7, and  
>> modifying a crystal structure PDB file taken from the PDB, with no  
>> hydrogens.

> There does appear to be a problem specifically with united-atom  
> force fields. Both CHARMM and AMBER function correctly, with or  
> without -ignh. Using any GROMOS variant, with or without -ignh leads  
> to a problem.

That's not true, actually, as far as I can tell. I did a bit more  
testing, and ASPH in the input does NOT get protonated for AMBER99 or  
CHARMM27 either. I tried with 2018.1.

The bug I reported earlier (assuming then it was Gromos only):  
https://redmine.gromacs.org/issues/2480