From: Peter Freddolino (
Date: Wed Jun 06 2007 - 10:41:54 CDT

Hi Himanshu,
your best bet is to make use of trans bond (as mentioned in my previous
email) on atoms of each residue separately; this way you avoid having
changes propagate further than you mean. For each residue just make a
selection with the atoms to be moved and use, for rotation about the
bond between atom1 and atom2,
$sel move [trans bond {atom1coord} {atom2coord} DISTANCE deg]
for both the phi and psi angles; you will have to have first measured
the starting coordinates for the phi and psi angles. You can obtain the
initial angle by making a label and then deleting it; eg.

label add Dihedrals $tmpmolid/$dihedatom(0) $tmpmolid/$dihedatom(1)
$tmpmolid/$dihedatom(2) $tmpmolid/$dihedatom(3)
set dihedral [lindex [lindex [label list Dihedrals] end] 4]
label delete Dihedrals all


Himanshu Khandelia wrote:
> I want to confine the changes only to the residues in question (so that
> only the structure of the loop changes, and not that of the rest of the
> structure). I was doing this using:
> for {set i 22 } { $i < 26 } { incr i } {
> set sel [atomselect top "resid $i and name CA"]
> $sel set phi -60
> $sel set psi -45
> }
> So the {phi,psi} pair of only residues 22 through 26 changed. But the
> entire molecule downstream rotates because of the changes enforced on the
> dihedral angles of residues 22-25. I hope its clearer now ?
> Thank you,
> -himanshu
> ----------------------------
> Himanshu Khandelia, PhD
> Research Assistant Professor (Postdoc)
> MEMPHYS, Center for Membrane Physics:
> -----------------------------
> On Wed, 6 Jun 2007, Peter Freddolino wrote:
>> Hi Himanshu,
>> I think I may not understand exactly what you're after -- do you want to
>> rotate one entire helix about the dihedral angle, or confine your
>> changes only to the residue under consideration? In either case the
>> answer is more or less the same: make a selection containing the atoms
>> you want to move and use the "trans bond" command
>> ( with the
>> coordinates of the atoms at either end of the bond to set up a
>> transformation matrix to apply to the selected atoms.
>> Best,
>> Peter
>> Himanshu Khandelia wrote:
>>> Hi,
>>> My protein consists of two transmembrane helices connected by a
>>> periplasmic loop. I want to artificially extend one of the helices into
>>> the loop region. One way to do this is to alter the dihedral angles of the
>>> relevant amino acid residues. However, changing the (phi,psi) values of
>>> some of the loop residues to helix specifications (-60,-45) will distort
>>> the topology of the protein downstream of the altered region. Does someone
>>> know of a way to alter dihedral angles of ONLY the specific region of the
>>> protein, while keeping the positions of the rest of the atoms constant ?
>>> Any other suggestions?
>>> Thank you for the help,
>>> ----------------------------
>>> Himanshu Khandelia, PhD
>>> MEMPHYS, Center for Membrane Physics:
>>> ----------------------------