Hi Grange,
I'll try this out on my machine and see if I get the same problem. However,
a few things to note. Firstly this may not be an infinite loop it may just
be running very very very slowly. Your system appears to be 190 atoms which
is pushing the limits for what mopac 6.0 was ever designed to do, in fact
there may be hard wired arrays in the code that are big enough to cope with
such a large system. In fact just looking at your mol2 file I count 103
heavy atoms which is 3 larger than the 100 max heavy atoms that mopac 6 is
configured for (see $AMBERHOME/src/mopac6/src/SIZES. In fact the run I just
tried on my machine segfaulted during the mopac run for likely the above
reason.
Antechamber 1.27 used Divcon for the charge derivation which is newer and
likely more capable of dealing with such large systems.
You could try increasing the following in SIZES:
PARAMETER (MAXHEV=100, MAXLIT=100)
to 150 each and recompiling and see if that helps.
That said neither case is probably ideal. Doing charge fits on such large
systems is really not recommended, there are all sorts of issues with buried
atoms, multiple local minima etc. In fact in your case your system comprises
a number of individual residues so you probably shouldn't be using
antechamber at all and it is not clear to me why you want to do it this way.
Except for the CHU, AIB and CHV everything looks to be standard residues:
1 GLU356 2 RESIDUE 4 B GLU 1 ROOT GLU B 356
2 GLU357 11 RESIDUE 4 B GLU 2 **** GLU B 357
3 VAL358 20 RESIDUE 4 B VAL 2 **** VAL B 358
4 ASP359 27 RESIDUE 4 B ASP 2 **** ASP B 359
5 CHU360 35 RESIDUE 4 B CHU 3 **** LEU B 360
6 GLU361 43 RESIDUE 4 B GLU 2 **** GLU B 361
7 ASP362 52 RESIDUE 4 B ASP 2 **** ASP B 362
8 AIB363 60 RESIDUE 4 B AIB 2 **** TYR B 363
9 CHV364 72 RESIDUE 4 B CHV 3 **** VAL B 364
10 ALA365 79 RESIDUE 4 B ALA 2 **** ALA B 365
11 ARG366 84 RESIDUE 4 B ARG 2 **** ARG B 366
12 PRO367 95 RESIDUE 4 B PRO 2 **** PRO B 367
13 ASP368 102 RESIDUE 4 B ASP 1 **** ASP B 368
So why would you want to use AM1-BCC charges and GAFF over a specific
amino-acid force field such as FF99SB? The non-standard residues look like
amino-acid type derivatives so you might be better off just looking for
analogues in the FF99SB force field that you can use as templates. I would
then manually do HF-6-31G* RESP fits yourself using ACE/NME capped versions
of each individual residue using the same approach as the original FFXX
force fields.
Alternatively if you really want to use Antechamber and AM1-BCC charges then
you should probably split this system up into individual residues and cap
them and run the charge fits individually.
All the best
Ross
From: owner-amber.scripps.edu [mailto:owner-amber.scripps.edu] On Behalf Of
Grange Hermitage
Sent: Sunday, July 13, 2008 6:07 PM
To: amber.scripps.edu
Subject: AMBER: AmberTools1.1 (and 1.0) antechamber endless loop
Hi,
I am getting an endless loop with AmberTools1.1 and 1.0 installations on
different machines. I have tested this out to 1 hour elapsed time.
However, with AnteChamber-1.27, the command completes and does so in 8
minutes.
The command is:
antechamber -fi mol2 -fo prepi -i chp3.mol2 -o chp3.prepin -j 5 -c bcc -s 2
The chp3.mol2 is attached.
Perhaps one of the developers could execute this in their AmberTools
environment to see what result they get there.
I have a tgz of the directory used for the run, and will send if anyone
wants to see that.
Is a different means of reporting a possible bug recommended?
Thanks,
Grange.
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
to majordomo.scripps.edu
Received on Wed Jul 16 2008 - 06:07:13 PDT
