[gmx-users] g_hbond problem to select donor atoms in drug

[Mark Abraham]

geeta kant wrote:
> 
> 
> Dear Gromacs users,
> I am trying to analyze hydrogen bonding between a Protein and Drug by 
> g_hbond from .xtc and .tpr and .ndx files. I am getting expected results 
> when the H-bond donor and the hydrogen atom is from protein while the 
> hydrogen bond acceptor atom is in the drug. The results are, however, 
> wrong when I am trying to look for H-bond when the donor and hydrogen 
> atoms are from the drug molecule and acceptor atoms are in the protein.  
> My atom names in the drug molecule sometimes start with N or A, such as 
> NC6 (a carbon), NN6 (a nitrogen), NH61 (a hydrogen) etc. and g_hbond is 
> considering all these atoms as hydrogen bond acceptor.  I would 
> appreciate if somebody  gives a hint to calculate the correct hydrogen 
> bonds in my system.

Apparently the heuristic for detecting H-bonding involves the atom 
names, and ones starting with N are guessed to be nitrogen. I suggest 
going back to your pre-grompp coordinate file, editing the atom names 
suitably, doing the corresponding edits to the atom names in the [ 
molecule ] section(s) of your topology. Then you can regenerate a new 
.tpr with atom names with which g_hbond will work better. So long as the 
order of the atoms has not changed, you can combine your new .tpr and 
your old .xtc as inputs to g_hbond.

Mark