thanks for your help, it is difficult for me to imagine the effect of this operation, but I have read a few places that this should be the effect, but I might be wrong. Anyway, what I want is using two consecutive C-alphas as "flexible hinges" rotating the peptide plane between this two C-alphas around the imaginary bond between them leaving the coordinates of the rest of the molecule unchanged. How would I accomplish this?

best, Jakob

2010/6/28 Robert Campbell
Dear Jakob,

On Mon, 28 Jun 2010 15:32:59 +0200 Jakob Nielsen <jtoudahl.nielsen@gmail.com>
wrote:

> I would like to modify a protein pdb file with a "crankshaft" flip, which
> is the anti-correlated double change: psi(i-1) += delta and phi(i) -= delta.
> Such a change should leave the protein coordinates unchanged effecting only
> atoms in residues i-1 and i. However implementation in pymol (see below)
> changes the coordinates of _all_ residues from i-1 to the C-terminal. Can
> you help me? Can I somehow do the two changes simultanious?...
>
> set_dihedral 21/n, 21/ca, 21/c, 22/n,  148.435 #was 168.4
> set_dihedral 21/c, 22/n, 22/ca, 22/c, -83.704  #was -103.7

I don't see how this change can leave the other coordinates unchanged.  Even
if the psi(i-1) and phi(i) bonds are exactly parallel (meaning that the
angles CA-C-N and C-N-CA are exactly the same, which they are not
necessarily), rotation of the psi(i-1) bond will move the N(i) and
CA(i) atoms, so therefore while a rotation in the opposite direction of the
phi(i) bond will tend to keep the rest of the change travelling in the
direction it was before those two counter-rotations, the backbone itself will
be shifted.  One textbook I consulted lists the angles at the C(i-1) and N(i)
atoms as 116 and 122, respectively, so the psi(i-1) and phi(i) bonds are not
even theoretically going to be parallel.  In practice, with "real"
structures, they certainly are not usually going to be parallel.

Cheers,
Rob

--
Robert L. Campbell, Ph.D.
Botterell Hall Rm 644
Department of Biochemistry, Queen's University,