Re: [AMBER] question about dihedral angles

From: Karl Kirschner <>
Date: Wed, 2 Dec 2020 08:53:32 +0100

Hi, Richard,

  You can rotate a dihedral in tleap by using the "impose" command. Let's
say you have an tleap object named "mymolecule" that is composed of several
residues, and you want to rotate residue 7's atoms that are named C1, C2,
C3 and C4 to 120 degrees, then the following command should work:

impose mymolecule {7} { {C1 C2 C3 C4 120.0} }

Note that it becomes a bit unclear if you have rotating a dihedral is
between two residues. Then you indicate the lower residue number, and start
the sequence with it's atom names. In the Amber20's manual, you can find
additional examples on page 242 (subsection entitled "Procedures for
building oligosaccharides using the GLYCAM-06 parameters").

An alternative approach is to use PyMol, which would enable you to visually
see the rotation being made and if any atomic clashes happen. To do this
you need to label the atoms via their "ID" and then use the command
"set_dihedral" (e.g. "set_dihedral ID 6, ID 7, ID 8, ID 9, 120") Using
PyMol will add a few steps to your workflow since the resulting structure
must be saved to a PDB-formatted file and then be loaded back into tleap.
However, this approach will help maintain a precise understanding of what
is structurally happening to your molecule with your manual manipulation.
Furthermore, you can create a PyMol script that allows you to encode these
changes and then have a reminder of exactly what you did for your model
building, increasing the reproducibility of the workflow if needed later.


On Wed, Dec 2, 2020 at 3:07 AM David A Case <> wrote:

> On Tue, Dec 01, 2020, Gordon Richard Chalmers wrote:
> >
> >I have a quick question as to how AmberTools20, tleap,
> >or cpptraj rotates the molecule about a phi or psi dihedral
> >angle.
> The world's experts on this subject are right in your backyard: check
> with Rob Woods or Lachele Foley, who probably have the longest
> institutional memory about this.
> The basic problem is that just saying you want to rotate about an angle
> is not enough information. One also needs to specify which atoms (i.e.
> which "side" of the torsion) should be moved. But the tleap syntax
> doesn't include this, and it can be rather difficult to understand its
> internal procedures for making this decision.
> I'm less familiar with cpptraj handles this problem.
> ....dac
> _______________________________________________
> AMBER mailing list
AMBER mailing list
Received on Wed Dec 02 2020 - 00:00:02 PST
Custom Search