AMBER Archive (2004)

Subject: AMBER: Bad contacts

From: bybaker_at_itsa.ucsf.edu
Date: Mon Jun 07 2004 - 19:44:37 CDT

Hello, Amber:

I built a protein model from the homology crystal structure using
Swiss-3D modeling program. Hydrogen atoms were added to this model in
tleap. And the model was subjected to sander (Amber7) energy
minimization under non-periodic boundary condition. The idea is to have
a simple run for the entire molecular system. It went through.

The total energy level was lowered from ‘2.1754E+03’ to ‘-3.1382E+03’;
RMS was lowered from ‘2.1393E+02’ to ‘4.4849E+00’; VDWAALS was lowered
from ‘4159.9961’ to ‘-827.5665’.

My input file for sander run is :

‘initial minimization test 1d
  &cntrl
    ntb=0, imin=1, maxcyc=200, ntpr=5
  &end’

Then I checked the structures of before and after energy minimization
using SwissPDB Viewer. For the initial structure, no bad contact was
found for either ‘aa making clashes’ or ‘aa making backbone clashes’.
For the structure after minimization, several residues are identified in
making clashes and making backbone clashes. I notice that there are some
complaints when I loaded the minimized structure in SwissPDB Viewer, such
as ‘some hydrogen atoms could not be connected’, ‘atoms of unknown group
have been treated as HETATM. Numbers of groups treated as HETATM (5)’,
and ‘’at least one HETATM group lacked proper CONNECT information,
connections will be generated between atoms that are closer than 2.000A,
which can generate false bonds. Nb of HETATM groups in this case (10)’.

My questions are:

(1) The bad contacts in structural (after minimization) are caused by
       
<a> the program incompatibility between Amber and SwissPDB Viewer? Such
as HIS in SwissPDB Viewer while HID in Amber. When loading pdb file in
tleap, Amber automatically add the missing atoms to the file. These atoms
may be the trouble fro SwissPDB Viewer to recognize?: or by

     
<b> the different on force field definition between SwissPDB Viewer and
Amber7. When I ran tleap (Amber7), I used .ff99 for loading .pdb files.

(2) Are there any improper setting in my input file for sander run? If
so, is this the reason for these bad contacts? If not, how energy
minimization create some bad contacts? While it suppose to improve the
structure modeling as I understand.

I am bit of confused on which structure model should be used. Your help
and advice will be much appreciated.

Thank you

Bo Yang

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