From: Gumbart, JC (
Date: Tue Aug 03 2021 - 14:44:09 CDT

Yeah, we used "vpbonds 0” for the sake of NAMD. The exact error we get is

Info) Using plugin psf for structure file nWater.psf
psfplugin) Failed to parse atom line in PSF file:
psfplugin) ' 1 WT1 5 SWM4 OH2 ODW 1.716386 15.5994 0 -0.978253 1.300000'
couldn’t read atom 0

This is with VMD194a38, but a48 produced the same. We addressed it for now by manually converting the spacing so that it’s an EXT-formatted PSF.


On Aug 3, 2021, at 3:38 PM, Joao Ribeiro <<>> wrote:

Hi JC,

I am not sure about the error that you are getting, but the only difference that I can remember between NAMD and VMD is the explicit bond between the lone pair/Drude particles and the host atom. NAMD requires that these bonds are omitted, but for VMD to show the bonds, one needs to run psfgen with vpbonds 1

vpbonds [1 0]
Purpose: Print the bonds between the virtual particles (drude particles and lone pairs) and their
Arguments: 1 for true, 0 for false. Default value 1.
Context: Before writing the psf file. May call multiple times. WARNING: To run simulations containing lone pairs or Drude particles on NAMD 2.13, set vpbonds to 0

It's been a while since I touched this code, so forgive me if there is anything else that I missing.



On Tue, Aug 3, 2021 at 3:23 PM Gumbart, JC <<>> wrote:

Psfgen 2.0 supports Drude, and we used it recently to build a simple water box. NAMD can use the resulting files just fine, but VMD throws an error loading the PSF. It expects it to be in the EXT format but the file produced by Psfgen is not (no long names or anything that would normally trigger it I guess). We can fix this ourselves in the meantime, but generally, should Psfgen be making Drude PSFs EXT by default? Or is it an issue with VMD’s expectations?

Or am I misunderstanding something?