Re: Help! unstable simulation - ARG go nuts...

From: Jim Phillips (
Date: Tue Dec 16 2003 - 22:10:31 CST


It looks like this is your problem:

unknown patch type PT1
ERROR: failed to apply patch

(Yes, psfgen should really crash and burn here...)

I think you meant to type TP1. If you do, you get:

applying patch TP1 to 1 residues
unknown atom type ON3
add atom failed in patch TP1
ERROR: failed to apply patch

Your top_PTR.inp file contains the following note:

! this patch requires that the use of the following
!1) protein topology file that contains mass list of atoms in both
! the protein and nucleic acid topology files and
! add to the topology file
!MASS 179 ON2b 15.999400 ! Nucleic acid phosphate ester oxygen (pres
!2) parameter file that contains both the protein and nucleic acid
! parameters plus the parameters listed below

If I substitute the prot_na topology file for prot_lipid there are other
atom types missing, but if you include them both then psfgen is happy.

Hopefully this will fix your minimization problems.


On Tue, 16 Dec 2003, Gadi Oron wrote:

> Hello,
> On Fri, 2003-12-12 at 11:16, Alexandre Vakhrouchev wrote:
> Thank you for your reply.
> >
> > First of all look at the output of your psfgen script (it is attached
> > as 'pfs.log').
> I have noticed that some coordinates are guessed, but I don't think that
> this is the reason for the crash. You are right though that the the
> crash is due to close atoms, I have written a script that finds atoms
> too close to each other, and found that at the end of the minimization
> some atoms are literally on top of each other. The problem is that this
> was not the case before the minimization! ARG residues are especially
> bizarre. Their atoms simple collapse during minimization to distances
> less than half an A(!!). I have attached a snapshot to illustrate it.
> I do believe that this in NOT due to the patches that I apply since
> before minimization everything is fine. I have also isolated a single
> chain that is not patched or modified, and it behaves the same (chain
> C).
> The only way I could get the dynamics running is to specify
> "exclude none", but in this case the protein swells into unrealistic
> dimensions (H bonds become 2A long!)...
> Any ideas?
> > I solved this error by providing coors for all atoms,
> > depending on Rmin/2 values (from Van-der-Waals force field
> > parametres), but i had to
> > write additional programm.
> I would be glad to have this prog... even though I doubt it would be of
> great help...
> Thank you for your help.
> --
> ================= Gadi ORON, PhD -o- Proteologics ======================
> =========== gadi/at/ =========== +972 8 9475666 ===========
> Also, the Scots are said to have invented golf. Then they had
> to invent Scotch whiskey to take away the pain and frustration.

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:37:14 CST