RE: Re: vmd-l: PSFGEN 1.6.4 autogenerate PATCH keyword not available

From: Vermaas, Joshua (
Date: Tue Apr 24 2018 - 12:08:20 CDT

Yes. I think the way it is implemented in CHARMM is that angles and dihedrals are automatically regenerated after every patch command, or perhaps are deferred until writing the psf rather than immediately after segment generation.

On 2018-04-24 10:23:51-06:00 Brian Radak wrote:

Oh I think I see now. This line isn't within a residue or patch definition, it's in the RTF "header". I'm still not clear what this is supposed to mean, though. Is it to indicate that the "first", "last", and "patch" commands should be understood to imply the coda "regenerate angles dihedrals"?

I think the best we can implement in psfgen is that these commands are ignored, much like the CHARMM NONBOND commands. In fact, it's not really clear that psfgen would do anything useful with that command even if it were correctly parsed (why would it regenerate while loading topology definitions?).

On Tue, Apr 24, 2018 at 11:59 AM, Vermaas, Joshua <<>> wrote:
Basically psfgen shouldn't throw an error when it encounters "PATCH" in the default behavior it sees in a standard CHARMM .top file. Right now, in CHARMM, angles and dihedrals can be automatically regenerated after a patch is applied if that flag is present in the .top file. psfgen doesn't recognize it, and instead psfgen expects a "regenerate angles dihedrals" command after patches have been applied.


On 2018-04-24 06:39:23-06:00<> wrote:

Hi Tom,

The "segment" command supports the special patch keywords "first" and "last" but the "patch" command is expected as a post-segment modification only.

What usage pattern were you expecting here? It seems weird that a residue definition would only make sense after a patch has been applied.


On Mon, Apr 23, 2018 at 8:57 PM, Thomas C. Bishop <<>> wrote:
(i'm sending to both since vmd and namd users may independently be using psfgen... this was also reported in VMD-l about a year ago but does not seemed to have changed.

THe post was originally
"From: Ashar Malik (<>)
Date: Wed Apr 05 2017 - 05:15:38 CDT"

It seems that psfgen autogenerate does not recognize the "PATCH" option for the AUTO keyword
as evidenced by
vmd > auto
arguments: ?angles? ?dihedrals? ?none?

so line 92 in CHARMMS top_all36_prot.rtf file that reads

    throws an error... presumably b/c psfgen does not know the "PATCH"option
    but the edited line
    works as it expected

It seems reasonable (and likely easy fix) that the AUTOGENERATE command should work on patches too.

Is this a bug or feature?
 output providedbelow

Thanks in advance for any info


*************************************************** SESSION DUMP **************************

Info) VMD for LINUXAMD64, version 1.9.3 (November 30, 2016)
Info) Email questions and bug reports to<>
Info) Please include this reference in published work using VMD:
Info) Humphrey, W., Dalke, A. and Schulten, K., `VMD - Visual
Info) Molecular Dynamics', J. Molec. Graphics 1996, 14.1, 33-38.
vmd > package require psfgen
topology toppar/top_all36_prot.rtf
vmd > psfgen) reading topology file toppar/top_all36_prot.rtf

psfgen) >>>>>>>>CHARMM36 All-Hydrogen Topology File for Proteins <<<<<<
psfgen) >>>>> Includes phi, psi cross term map (CMAP) correction <<<<<<<
psfgen) >>>>>>>>>>>>>>>>>>>>>>>>>> May 2011 <<<<<<<<<<<<<<<<<<<<<<<<<<<<
psfgen) All comments to the CHARMM web site:<>
psfgen) parameter set discussion forum
psfgen) Created by CHARMM version 36 1
psfgen) ERROR! Failed to parse autogenerate statement. Line 92: AUTO ANGLES DIHE PATCH

psfgen) cross-term entries present in topology definitions


This archive was generated by hypermail 2.1.6 : Mon Dec 31 2018 - 23:21:02 CST