AMBER Archive (2009)
Subject: Re: [AMBER] (no subject)
From: Jenny Iskrenova (jnova2001_at_gmail.com)
Date: Fri Jan 16 2009 - 17:20:46 CST
> Something is really wrong at this step, and you will need to track it
> down. Leap just uses the coordinates you give it. So if you load a
> "good" pdb, but the resulting inpcrd coordinates are "bad", something
> has gone wrong, and you should not go further with MD, etc.
> It might be that the problem is the interface to VMD -- does VMD look
> "good" with the "good initial pdb" file? You may have to actually
> examine the atomic coordinates in the inpcrd file (use ambpdb to convert
> to pdb) vs. those in the input file, to find out what is going on.
The pdb file, created with Materials Studio looks fine in VMD. This is the file:
REMARK Materials Studio PDB file
REMARK Created: Fri Dec 12 20:28:51 Eastern Standard Time 2008
ATOM 1 C1 VIN 1 -2.841 -0.305 0.052 1.00 0.00 C
ATOM 2 C2 VIN 1 -1.773 -0.445 1.101 1.00 0.00 C
ATOM 3 H1 VIN 1 -2.694 -1.066 -0.730 1.00 0.00 H
ATOM 4 H2 VIN 1 -2.795 0.694 -0.410 1.00 0.00 H
ATOM 5 H3 VIN 1 -3.838 -0.444 0.501 1.00 0.00 H
ATOM 6 O1 VIN 1 -0.712 0.194 0.958 1.00 0.00 O
ATOM 7 O2 VIN 1 -1.986 -1.198 2.071 1.00 0.00 O
I have changed the default MOL to VIN and made the atom names unique
i.e. instead of just C and C, I made them C1 and C2, etc.
> Do the simplest possible LEap run (no addions, no solvent, just make a
> vaccum model). See if things are getting distorted. Be sure to look
> carefully at any warnings LEaP gives (missing atoms, extra atoms, etc.).
> It is certainly possible that leap doesn't understand the Materials
> Studio pdb file, for example.
This is exactly, what I did. First, I ran antechamber to get the prepi
antechamber -i acetate.pdb -fi pdb -o acetate.prepin -fo prepi -c bcc
-s 2 -nc -3 -j 5
Then, this is the simple script that I run with tleap:
saveoff VIN acetate.lib
v1 = loadpdb acetate.pdb
saveamberparm v1 acetate-vac.prmtop acetate-vac.inpcrd
The only warning that tleap gives is about the net charge of the
molecule being non-zero. The structure looks fine in xleap, as well.
The bonds look distorted when I load acetate-vac.prmtop and
acetate-vac.inpcrd in VMD.
Finally, I restored the acetate-vac.pdb file using ambpdb and
acetate-vac.prmtop and acetate-vac.inpcrd files. The coordinates of
the corresponding atoms are the same. The only difference is the order
of the atoms, e.g. C2 is now atom #5 instead of atom #2 in the
original pdb file. This is the new pdb file:
ATOM 1 C1 VIN 1 -2.841 -0.305 0.052
ATOM 2 H1 VIN 1 -2.694 -1.066 -0.730
ATOM 3 H2 VIN 1 -2.795 0.694 -0.410
ATOM 4 H3 VIN 1 -3.838 -0.444 0.501
ATOM 5 C2 VIN 1 -1.773 -0.445 1.101
ATOM 6 O2 VIN 1 -1.986 -1.198 2.071
ATOM 7 O1 VIN 1 -0.712 0.194 0.958
I have attached a pdf file with the distorted structure generated by VMD.
I tried loading the structure (acetate-vac.inpcrd) with and without
periodic box but it did not make any difference.
Would this, really, be a problem with VMD?
- application/pdf attachment: vmd.pdf
AMBER mailing list