On Thu, Sep 24, 2015 at 8:48 AM, David A Case <david.case.rutgers.edu>
wrote:
> On Thu, Sep 24, 2015, Damiano Spadoni wrote:
> >
> > | ERROR: PMEMD does not support intermolecular PRFs!
> >
> > As you suggested, I checked my topology file with checkValidity (this
> > kind of errors may due to something wrong in the .prmtop file) and it
> > detected two CYS residues being within 3A distance and it suggested to
> > change the name of the residues and build a disulfide bond, without
> > specifying which cysteines it is referring about.
>
> Well, you should certainly ignore the "suggestion" about disulfide
> bonds....
>
> But if you have an SF4 resiude covalently connected to your protein, that
> could likely be the source of the problem pmemd reports. My understanding
> is
> that parmed should be able to fix this, but Jason knows more about this
> than
> I do.
>
​ParmEd should be able to fix issues in which molecules are non-contiguous,
but checkValidity should have issued a warning about that.
What I currently suspect as happening right now is that not all atoms in
the Fe4-S4 cluster are connected to each other through bonds. This is what
I was going to check when Damiano sent me his topology file. If you *know*
that this is true (i.e., that there are distinct "molecules" defined within
the Fe4-S4 cluster), then this is your problem. All atoms within a residue
are *required* to be part of the same "molecule" by pmemd's parallel
decomposition scheme. Which means you would need to break up your residue
so that the previous criteria is satisfied.
If you do not think that this is true, it needs to be checked (which is
what I was going to do).
HTH,
Jason
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Sep 24 2015 - 06:30:03 PDT
