From: Vermaas, Joshua (
Date: Mon Oct 14 2019 - 09:26:56 CDT

Hi Raul,

Digging around in the find_connected_fragments code within the VMD source, I found the issue. Normally, find_connected_fragments traverses the bonds in your system to assign fragments, with connected atoms with the same chain and segname being assigned to the same fragment. However, for reasons that I'm sure made sense at the time, there is an exception for otherwise connected components joined by a disulfide (atom names are SG), whose bonds are not considered when going through this joining process. So, if folks in the future finding this through google want to prepare something in topogromacs (which uses fragments to tell what is connected or not) for a system where disulfides link together the system, try something like this:

set disu [atomselect top "name SG"]
set sel [atomselect top "all"]
#Make everything the same segname and chain
$sel set segname "SEG"
$sel set chain A
#Change the name of the atoms in the disulfide so that the bond traversal realizes that this is one connected fragment, which is how gromacs molecules are defined
$disu set name SD
#The reanalyze command will go through the bondlist and regenerate fragment definitions
mol reanalyze top
#Change the names back after reanalyzing. Does not change anything
$disu set name SG
topo writegmxtop [list parameter files go here]


On 2019-10-14 08:00:24-06:00 wrote:

Dear VMD users:
I have a protein that consist on two separated polipeptide chains connected by a disulphide bond. I want to join them a a unique fragment to process it with writetopgmx (topotools) to build the gromacs topology. I have tried changing the segname and chain name of the protein so that both chains have the same name and segname. However VMD still considers them as separated chains. I checked on the psf that the inter-chain S-S bond is present and it is. How can I get VMD to understand that these chains are connected so they can be treated as one fragment?

Raul R Araya-Secchi
Postdoctoral researcher
X-ray and Neutron Science,
The Structural Biophysics Group,
Universitetsparken 5,
2100 CPH, DK.<>