Re: incompatibility between NAMD and amber topology files

From: Osman, Roman (roman.osman_at_mssm.edu)
Date: Tue Jan 10 2012 - 15:33:43 CST

Irene,
Thanks, I can try using sleap, but what you are describing could produce a lot of trouble.
Nevertheless I think that there is problem with reading amber topology files in namd, so someone should fix it. I think the change that was introduced into the amber topology files presents a greater flexibility of assigning the non bonded scaling parameters. It would be nice if name was able to incorporate it or at least read through it without causing an error.
Rami

On Jan 10, 2012, at 4:25 PM, Irene Newhouse wrote:

I've had this problem, too. I got around it by switching to sleap instead of tleap. There's a downside - sleap doesn't always properly neutralize my systems, so I have to do a quick trial NAMD minimization & I read the CHARGE entry in the log file to make sure it's 0. I look at the number of neutralizing ions in the system & adjust accordingly if the charge isn't essentially 0. tleap agives you the number of water molecules you added to create your box, while sleap gives you the number of protein residues & the total number of residues instead, but that's minor compared to not always neutralizing properly.

Irene

________________________________
From: roman.osman_at_mssm.edu<mailto:roman.osman_at_mssm.edu>
To: namd-l_at_ks.uiuc.edu<mailto:namd-l_at_ks.uiuc.edu>
Subject: namd-l: incompatibility between NAMD and amber topology files
Date: Mon, 9 Jan 2012 22:13:24 +0000

Greetings,
I am not sure whether I should post it to the amber mailing list or the namd. Since I created the topology file with leap I sent the message to the amber mailing list, but they suggested that I also send it to the NAMD list.

I wanted to bring to your attention that NAMD cannot read a topology file created by tleap in amber.
I prepared a regular topology file for a protein in a rectangular box using the ff99SB. NAMD complained with this message:

Reading parm file (k7lp_r.top) ...
PARM file in AMBER 7 format
FATAL ERROR: Failed to read AMBER parm file!
[0] Stack Traceback:
  [0] CmiAbort+0x5f [0xabb81b]
  [1] _Z8NAMD_diePKc+0x62 [0x509262]
  [2] _ZN9NamdState14configListInitEP10ConfigList+0xcad [0x90bab7]
  [3] _ZN9ScriptTcl9initcheckEv+0x5e [0x9753e0]
  [4] _ZN9ScriptTcl12Tcl_minimizeEPvP10Tcl_InterpiPPc+0x20 [0x971f9c]
  [5] TclInvokeStringCommand+0x91 [0xae0518]
  [6] /share/apps/NAMD_2.7b2/namd2 [0xb16368]
  [7] Tcl_EvalEx+0x176 [0xb169ab]
  [8] Tcl_EvalFile+0x134 [0xb0e3b4]
  [9] _ZN9ScriptTcl3runEPc+0x14 [0x9714da]
  [10] _Z18after_backend_initiPPc+0x22b [0x50da7b]
  [11] main+0x3a [0x50d81a]
  [12] __libc_start_main+0xf4 [0x301941d994]
  [13] _ZNSt8ios_base4InitD1Ev+0x4a [0x508bda]

Comparison of the current topology file with one that worked before finally focused on two additional sections that were not there before; these are:
%FLAG SCEE_SCALE_FACTOR and
%FLAG SCNB_SCALE_FACTOR

When I removed those, NAMD read successfully the topology file and is running so far OK. I did not check the energies, but the behavior is of a normal NAMD job.

Perhaps it could be useful to introduce a fix in NAMD so it could read amber topology files without the need to manipulate them.

Thanks for your help,
Rami

Roman Osman
Department of Structural and Chemical Biology
Mount Sinai School of Medicine
New York, NY 10029
Tel: 212-659-8627
Fax: 212-849-2456
e-mail: roman.osman_at_mssm.edu<mailto:roman.osman_at_mssm.edu>
 http://atlas.physbio.mssm.edu/~osmanlab>

Roman Osman
Department of Structural and Chemical Biology
Mount Sinai School of Medicine
New York, NY 10029
Tel: 212-659-8627
Fax: 212-849-2456
e-mail:
roman.osman_at_mssm.edu<mailto:roman.osman_at_mssm.edu>
 http://atlas.physbio.mssm.edu/~osmanlab

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:21:07 CST