Hi,
On Tue, Aug 08, 2017 at 05:22:54PM -0400, Scott Brozell wrote:
>
> On Tue, Aug 08, 2017 at 12:51:44AM +0000, Tsai Kun-Lin wrote:
> >
> > I am using AmberTools17 and try to parameterize a drug like molecule using antechamber and parmchk2 with gaff2. The followings are the commands I use:
> > antechamber -i Structure3D.sdf -fi sdf -rn LIG -o Structure3D.mol2 -fo mol2 -c bcc -nc 0 -pf yes -dr no -at gaff2
> >
> > parmchk2 -i Structure3D.mol2 -f mol2 -o Structure3D.frcmod -s 2
> >
> > The first command finished successfully, however, the second command returned me an error message and an empty frcmod file:
> >
> > Warning: Atom type (cs) is not in PARMCHK.DAT; using default values
> > for improper_flag [0], group_id [0], and equivalent_flag [0].
> > It is recommended to add the new atom type (cs) to PARMCHK.DAT.
> > Segmentation fault (core dumped)
>
> Using an up to date patched AmberTools 17, when i run the parmchk2
> command above on the mol2 file created by the first command it produces
> the attached Structure3D.frcmod without complaint.
>
> In addition, valgrind does not report a problem.
>
> You could rebuild antechamber with debugging symbols and report the trace:
> cd AmberTools/src/antechamber/
make clean
> make AMBERBUILDFLAGS='-g'
> parmchk2 -i Structure3D.mol2 -f mol2 -o Structure3D.frcmod -s 2
>
> And you could investigate your AmberTools installation:
> ./update_amber --show-applied-patches
> cd AmberTools/test/antechamber/
> make
> make check
I was wrong that i was using a fully patched AmberTools17
(i was using a fully patched Amber16). I can reproduce the warning
in a fully patched AmberTools17:
./parmchk2 -i Structure3D.mol2 -f mol2 -o Structure3D.frcmod -s 2
Warning: Atom type (cs) is not in PARMCHK.DAT; using default values
for improper_flag [0], group_id [0], and equivalent_flag [0].
It is recommended to add the new atom type (cs) to PARMCHK.DAT.
but it does not crash and produces a (different) Structure3D.frcmod.
I may be able to investigate this further tonight...
scott
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Aug 08 2017 - 20:30:03 PDT