AMBER Archive (2004)

Subject: Re: AMBER: resp error message

From: Andre Farias de Moura (andre_at_qt.dq.ufscar.br)
Date: Wed Sep 29 2004 - 12:41:05 CDT

hi Francois,
in fact I used XRED to do the resp fitting. I defined
only one orientation in my reference pdb file:

REMARK REORIENT 38 71 73
REMARK
ATOM 1 CT1 PPEEB 20.926 9.330 13.547 1.00 0.00 C
ATOM 2 H1 PPEEB 21.879 9.323 14.073 1.00 0.00 H
ATOM 3 H1 PPEEB 20.982 10.018 12.704 1.00 0.00 H
ATOM 4 H1 PPEEB 20.138 9.652 14.227 1.00 0.00 H
ATOM 5 CT2 PPEEB 20.631 7.995 13.063 1.00 0.00 C
ATOM 6 H2 PPEEB 20.576 7.308 13.906 1.00 0.00 H
ATOM 7 H2 PPEEB 21.419 7.674 12.383 1.00 0.00 H
ATOM 8 CT2 PPEEB 19.362 8.005 12.362 1.00 0.00 C
...

as to the number of restraints, I defined them in the pdb file, which
RED used to generate input1, but I don't think there were too many,
they were only 4 oxygen atoms out of 112 atoms. I also tried to write
my own input file to run the resp program. I tried to fix the charges
of aliphatic atoms to their typical values while the aromatic moiety
charges were optimized, but then I got 'nan' message for all the
unconstrained atomic charges.

can you figure out what could be wrong?

thanks again,

andre'

On Wed, 29 Sep 2004, FyD wrote:

> > I'm using resp to generate the atomic charges of an
> > organic molecule but I'm getting the following error
> > message:
> >
> > Number of unique UNfrozen centers= 106
> >
> > Non-linear optimization requested.
> > chgopt: LU decomp gave almost-singular U
>
> I got the same problem. If you use several molecular orientations try to
> decrease the number of orientations... It is related to the number of
> restraints you use in your input1... Francois
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
>

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu