[gmx-users] g_hbond -hx bug ?

[Marie-Pierre DURRIEU]

Dear all,

I have a problem with the option -hx of g_hbond in  the
3.3.99_development_20060306 version of GROMACS.
Indeed I have a 75-residue long helicoidal protein, and I expected to
obtain between 40 and 50 hbonds between residues n,n+4. But g_hbond gave
me about 10 (n,n+4) hbonds, whereas more than 40 (n,n>6) hbonds were
found, which is abnormally high.

I have modified gmx_hbond.c at line 1922, by changing :

 resdist=abs(top.atoms.atom[id].resnr-top.atoms.atom[ia].resnr);

by :

 resdist=abs(top.atoms.atom[i].resnr-top.atoms.atom[j].resnr);

It seems to me that id and ia refer to donor and acceptor indexes, and not
to the main index required by top.atoms.atom.`
With this modification I obtain more sensible results.

As I'm not a specialist in C, I may be wrong. So I would like to know if
my modification makes any sense.

Thanks in advance,


Marie-Pierre.