On Mon, Jul 02, 2012, case wrote:
> On Mon, Jul 02, 2012, Ilyas Yildirim wrote:
> >
> > My question is what this (bad) prmtop file will do on the pressure
> > calculations. In both, (bad) prmtop and (right) prmtop files, I have
> > exactly the same ATOMS_PER_MOLECULE sections. Does that mean the pressure
> > calculations will be the same, as well? Or during the pressure
> > calculations, some other sections in prmtop files (combined with
> > ATOMS_PER_MOLECULE) are used? Thanks.
>
> My guess is that the pressure errors will be completely negligible. Assuming
> you have lots of water molecules around your biomolecules, the water is the
> dominant contribution to the pressure. You could do a short re-run of some
> segment to check this.
Sorry...I think I misread your original questions, and have lost the thread
of the argument. If the two prmtops have the same ATOMS_PER_MOLECULE section,
then I think the pressures should be the same. But, if the have the same
ATOMS_PER_MOLECULE section, then ambpdb output should also be the same. So
I'm kind of lost here. [It's too late at night....]
Thinking back, I think I misunderstood what the real problem was with TER
cards. If your metal ions are bonded to both chains, to create one giant
supermolecule, then ambpdb will just put a single TER card at the end of the
supermolecule. Is that what you are seeing but don't want?
Currently, for periodic systems, ambpdb puts a TER card at the end of each
real molecule. Perhaps more like what Amber needs (and maybe what the PDB
intended) is that there should be a TER card if there is no bond between
adjacent residues. (Even this algorithm can be fooled if the users adds a
bond between a metal ion and the residue right before it, but right now,
I can't think of an easy way to detect such a situation.)
You can remove the section between lines 436 and 472 of ambpdb.F90 to get
the behavior described in the previous paragraph. I'll probably change things
this way going forward.
....dac
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Mon Jul 02 2012 - 19:00:03 PDT
