Hi Dan et al,
So the issue here appears to be that in the case of how we are recommending
people build lipids using the Charmm lipid builder leap ends up not even
writing an ATOMS_PER_MOLECULE section. E.g. if you take the pdb from the
Charmm GUI Builder with water already present, load that into leap and use
the setbox command to make it periodic then the resulting prmtop has NO
ATOMS_PER_MOLECULE entry. This seems to work fine though from what Ben has
told me. You can run NPT for example. So I am not sure what the real
ATOMS_PER_MOLECULE section is actually for and it seems like we should not
be relying on it to be present or even correct. Only when you run the
solvatebox command does leap then write ATOMS_PER_MOLECULE to the prmtop.
The TER cards should be added based on a traverse of the bond list rather
than the ATOMS_PER_MOLECULE. Otherwise there are a substantial number of
issues that need fixing in leap before we can rely on it. For example, we
just came across another example where we build chlorophyll from a mol2 that
went through antechamber. The Magnesium is not bonded to the chlorophyll in
the first example but is in the same residue. This should be 112 atoms. We
then use solvate box and we end up with the atoms per molecule being
111,3,3,3... This works fine in pmemd for NVT but as soon as we switch to
NPT it quits since the sum of atoms per molecule does not match natom.
Hacking the prmtop to make 111 to 112 things then work BUT the question is
should this be 112,3,3,... or should it be 111,1,3,3,... - In which case
that magnesium should have been split by leap into a separate residue.
Eitherway the code in leap seems extremely flakey, it seems (anecdotally) to
be related to reading mol2 files rather than pdbs. So we either need to fix
leap (volunteers?) or agree to ignore ATOMS_PER_MOLECULE in the various
codes.
Comments?
All the best
Ross
> -----Original Message-----
> From: Daniel Roe [mailto:daniel.r.roe.gmail.com]
> Sent: Monday, May 21, 2012 4:09 AM
> To: AMBER Mailing List
> Subject: Re: [AMBER] lipid11 force filed question
>
> Hi,
>
> I think this issue may stem from a missing or incorrect
> ATOMS_PER_MOLECULE section in the prmtop. This information is used by
> cpptraj to determine where to print TER cards. You can tell cpptraj to
> explicitly generate this information by specifying the 'molsearch'
> keyword before loading the prmtop, e.g.:
>
> molsearch
> parm dppe.prmtop
> trajin dppe.inpcrd
> trajout dppe_new2.pdb pdb
>
> That should generate proper TER cards in the PDB.
>
> -Dan
>
> On Fri, May 11, 2012 at 12:24 PM, Tales Yuan <yukitenkou.gmail.com>
> wrote:
> > I use cpptraj but it also don't print TER cards.
> >
> > *cpptraj -i ptraj.in -p dppe.prmtop*
> >
> > 2012/5/12 Daniel Roe <daniel.r.roe.gmail.com>
> >
> >> Cpptraj should print TER cards in PDB files; it will take the same
> input
> >> as ptraj.
> >>
> >> -Dan
> >>
> >> On May 11, 2012, at 12:11 PM, Tales Yuan <yukitenkou.gmail.com>
> wrote:
> >>
> >> > Thanks Walker. I use Amber11 + AmberTools12 now. I set $AMBERHOME
> to
> >> > amber12 directory as the manual said, and set the amber12/bin
> before
> >> > amber11/bin in PATH.(Page 23)
> >> >
> >> > *echo $AMBERHOME*
> >> > */home/user/amber12*
> >> > *echo $PATH*
> >> >
> >>
> */home/user/amber12/bin:/home/user/amber11/bin:/usr/local/cuda/bin:/...
> ....*
> >> > *which tleap*
> >> > *~/amber12/bin/tleap*
> >> >
> >> > I tried ptraj and ambpdb to convert pdb from inpcrd/prmtop.
> >> >
> >> > *ptraj dppe.prmtop ptraj.in* (dppe_new1.pdb)
> >> > *ampbpdb -p dppe.prmtop < dppe.inpcrd > dppe_new2.pdb*
> >> >
> >> > In new1 pdb, there is NO TER card in pdbfile, either lipids or
> water. ( I
> >> > tried a pdb including TIP3P water )
> >> > In new2 pdb, the TER cards seems good.
> >> > *PA 1*
> >> > *PE 2*
> >> > *PA 3*
> >> > *TER*
> >> > *PA 4*
> >> > *PE 5*
> >> > *PA 6*
> >> > *TER*
> >> > *...*
> >> > _______________________________________________
> >> > AMBER mailing list
> >> > AMBER.ambermd.org
> >> > http://lists.ambermd.org/mailman/listinfo/amber
> >>
> >> _______________________________________________
> >> AMBER mailing list
> >> AMBER.ambermd.org
> >> http://lists.ambermd.org/mailman/listinfo/amber
> >>
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> > http://lists.ambermd.org/mailman/listinfo/amber
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon May 21 2012 - 12:30:06 PDT
