[gmx-users] Unphysical minimization results inconsistent within simulation (Kimberly Bowal)

[Kimberly Bowal]

Thank you Justin.

I figured out the issue: my atomic indices were not consistent between my
MC output and those expected by my topology using GROMACS.  I'm still not
sure why some of the molecules seemed more affected (ie crumpled) than
others, but when I correct these indices I get the results I expect.

Kimberly


> You wouldn't, and "slightly" different bond lengths should be trivial to
> resolve by energy minimization.
>
> Are all molecules identical? How were they parametrized and validated?
> Does a single molecule minimize stably in vacuo? A dimer?
>
> -Justin
>
> --
> ==================================================
>
> Justin A. Lemkul, Ph.D.
> Assistant Professor
> Virginia Tech Department of Biochemistry
>
> 303 Engel Hall
> 340 West Campus Dr.
> Blacksburg, VA 24061
> jalemkul at vt.edu <https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users> | (540) 231-3129http://www.biochem.vt.edu/people/faculty/JustinLemkul.html
>
>
> Thank you Justin.  Neither my MD or MC simulations are using PBC so I don't
> think atomic overlap is the problem.
>
> To help see this, here are snapshots of the whole molecular cluster before
> (
> https://photos.app.goo.gl/dDUJ8f2s3MXyfTSw2) and after (
> https://photos.app.goo.gl/tGReLmIYVFD4038s1) minimization. As you can see,
> not all of the molecules are crumpled and crumpled molecules are found
> throughout the cluster.
>
> One thought is that the input molecule geometries from the MC code (which
> are showing the behaviour pictured above) have slightly longer bond lengths
> than the input molecule geometries that are minimised successfully.
> However if that were the issue, why would I see some molecules in the
> cluster affected and others not?
>
> Kimberly
>
>
>
> > > Hello,
> > >
> > > When minimizing a cluster of simple aromatic molecules, GROMACS will
> run
> > > but gives me unphysical results (some molecules are crumpled into a
> ball
> > > and the energies are NaN).  I have gone through the topology file and
> > can't
> > > find an issue, and when I use the same topology file with a different
> > > geometry file of a cluster of the same molecules (generated in a
> > different
> > > way), I have no issues.  I know this issue often is due to atomic
> > overlap,
> > > but I can't see this at all in my system.
> > >
> > > I believe this problem is my .gro input file, which I am getting from a
> > > Monte Carlo program, but I can't understand why it would cause an
> issue.
> > > One thing in particular that confuses me is that within the same
> cluster
> > of
> > > molecules (.gro file from MC program), some of the molecules are fine
> and
> > > others are messed up.
> > > To illustrate, here are snapshots of two molecules in the cluster
> (others
> > > not shown) before the minimisation: https://photos.app.goo.gl/
> > > AHcV4qagnSurbDT12
> > > and the same two molecules after minimisation: https://photos.
> > > app.goo.gl/91sLieiUuUB1SHkm1
> > >
> > > I'm happy to include the contents of any input files if that would be
> > > helpful - please let me know.
> >
> > Are molecules overlapping via PBC? Is your MC code PBC-aware? The
> > example you showed appears to manifest large displacements of both
> > molecules, suggesting that you may be using a very small box and the
> > molecules are actually overlapping, so one is fine and the other gets
> > trashed.
> >
> > -Justin
> >
> > --
> > ==================================================
> >
> > Justin A. Lemkul, Ph.D.
> > Assistant Professor
> > Virginia Tech Department of Biochemistry
> >
> > 303 Engel Hall
> > 340 West Campus Dr.
> > Blacksburg, VA 24061
> >
> > jalemkul at vt.edu | (540) 231-3129
> > http://www.biochem.vt.edu/people/faculty/JustinLemkul.html
> >
> > ==================================================
> >
>