From tripos!vela!victor@wupost.wustl.edu  Fri Apr 30 11:46:04 1993
Date: Fri, 30 Apr 93 16:46:04 -0500
From: tripos!vela!victor@wupost.wustl.edu (Vic Lewchenko)
Message-Id: <9304302146.AA13684@vela.tripos>
To: chemistry@ccl.net
Subject: RE: Multiple Molecule-Systems with Mopac?



In reply to Heiko Schinke's question concerning:

>Dear Subscribers,
>we are using MOPAC 5.0 (purchased with SYBYL 6.0). Recent we would compute
>a multi-molecule-system (see the *.dat-file below, generated with SYBYL).
>After some computation-time MOPAC stopped (during the computation for the 
>first reaction coordinate) and print in his output-file (see below):
>
>           THREE ATOMS BEING USED TO DEFINE THE
>           COORDINATES OF A FOURTH ATOM, WHOSE BOND-ANGLE IS
>           NOT ZERO OR 180 DEGREEES, ARE IN AN ALMOST STRAIGHT
>           LINE.  THERE IS A HIGH PROBABILITY THAT THE
>           COORDINATES OF THE ATOM WILL BE INCORRECT.
>
> and at the end: 
> 
>       ATOMS 65, 63, AND 54 ARE WITHIN 0.0059 ANGSTROMS OF A STRAIGHT LINE


   I would make the following suggestion: When setting up the MOPAC 
calcualtion in SYBYL, use the SETUP option Z_MATRIX to edit
the Z_MATRIX.  In SYBYL 6.0, you will have to do the setup from the
command line to use this option (it is not yet supported from the
dialog box).  At the command line enter:

SYBYL> QCPE M1 MOPAC SETUP <filename> Z_MATRIX

you will then see the z matrix listed to the screen.  Select row 66
to modify the definition of atom 66.  Enter the atom ids for the
atoms that will define the distance, angle and torsion internal
coordinates.  Examine the molecule on the screen and select atoms
which do not line up in a straight line.  Ideally the atoms you select
will be bonded to each other, but this is not always possible.  Also the 
atoms you select must be lower in id number than the atom you are 
defining (66).


I think this is the simplest solution to your problem.  Other possible
solutions would include renumbering your molecule or adding dummy atoms.
If you are still having problems, please e-mail me a mol2 file of your
molecule and I will be glad to try to help you further.


Regards,

Vic Lewchenko
Tripos Associates
St.Louis, MO  USA
E-mail: victor@tripos.com