VMD-L Mailing List

From: Tristan Croll (tristan.croll_at_qut.edu.au)
Date: Sun Jul 28 2013 - 18:54:19 CDT

Hi Thomas,

This has been discussed a few times in this list (I asked the same question about a year ago). Basically, the problem is that the long atom types break the rules of the original PDB standard, which is based on strict field lengths. Since VMD has to remain backward-compatible with standard PDB files it hasn't been modified to allow these yet, which leads to fairly random charges based upon whatever happens to fall within the columns designated as the charge field.

NAMD handles the long atom types with no troubles, though. The easy workaround, then, is to take your solvated protein before ionization, and run a brief NAMD simulation. In its startup, NAMD outputs a line similar to:

Info: TOTAL CHARGE = 3.13148e-05 e

Once it prints that, kill the simulation and calculate out manually how many ions you need to neutralise (and achieve your desired total concentration). Put those numbers into the relevant boxes under "User defined number of ions" in Autoionize. When you run, it will still (of course) spit out an error stating that the total system charge is non-zero, but it will give you a properly neutralised system.

Cheers,

Tristan

-----Original Message-----
From: owner-vmd-l_at_ks.uiuc.edu [mailto:owner-vmd-l_at_ks.uiuc.edu] On Behalf Of Thomas Albers
Sent: Sunday, 28 July 2013 7:37 AM
To: vmd-l_at_ks.uiuc.edu
Subject: vmd-l: Problem with autoionize plugin and long atomtypes

Hello!

There seems to be a problem when autoionize is used to add ions to a
system in which the "long atomtypes" from the CGenFF forcefield
appear.

When it reads in such a .psf file it calculates the initial charge as
non-integer, and then it throws an error message. Can anyone suggest
a workaround?

Regards,
Thomas