From: Justin Gullingsrud (jgulling_at_mccammon.ucsd.edu)
Date: Thu Feb 19 2004 - 12:28:52 CST

Hi John,

I think the problem is that VMD does not re-determine the protein
fragments and backbone connectivity of the molecule when you change the
bonds, even if you connect a C to an N. This means that ball-stick type
reps like Lines and Licorice will faithfully reflect the new bonds that
you assigned, but higher-level reps like Cartoon and Ribbons will not.
What you'll have to do for now if you want the higher level reps to work
is to modify the PSF file itself, either by putting that bond by hand
into the PSF file (shouldn't be hard to do, just remember to increment
the NBONDS number by 1), or by linking the segments together using psfgen.
There is a LINK patch residue expressly for this purpose.

Cheers,
Justin

On Tue, Feb 17, 2004 at 02:20:24PM -0600, Streiff, John H. wrote:
> I am preparing a fusion protein from two pdb files. I created a pdb/psf file of the two proteins correctly oriented in space, but as unconnected segments. I tried to use 'setbonds' to connect the two, but the command does not want to work for me when the two atoms are in different segments. (I can use setbonds to make a bons between atoms in the same segment). Any suggestion?
>
> This is the script I used (borrowed from an earlier posting):
>
> set sel [atomselect top "index 2000 2310"]
> set bonds [$sel getbonds]
> set ids [$sel get index]
> lassign $bonds atom1bonds atom2bonds
> lassign $ids atom1id atom2id
> lappend atom1bonds $atom2id
> lappend atom2bonds $atom1id
> $sel setbonds [list $atom1bonds $atom2bonds]
>
> - John