Re: [PyMOL] rms_cur

[Robert C. <rl...@po...>]

Jason,

* tree <jav...@ut...> [2005-05-12 11:39] wrote:
> PyMolers,
> 
> The docs leave me stranded here.  The wiki page
> (http://www.pymolwiki.org/index.php/Rms_Cur) is the same as the docs.
> 
> I'm trying to rms_cur two selections.  I keep getting an
>   ExecutiveRMS-Error: No atoms selected.
> error.
> 
> I load two proteins, 1gh2 and 1kao.  I then, created two easy selections
> A and D as:
> 
>   PyMOL>sel A, 1gh2 and n. CA and i. 65-99
>    Selector: selection "A" defined with 35 atoms.
> 
>   PyMOL>sel D, 1kao and n. CA and i. 64-98
>    Selector: selection "D" defined with 35 atoms.
> 
> So far, so good.

Actually not quite so good as you think...
 
>   PyMOL>rms_cur A, D
>   ExecutiveRMS-Error: No atoms selected.
> 
> Can anyone shed some light on this, please?

Yes, the problem is the lack of identity within your selections.

If you do "help rms_cur" it prints:

  DESCRIPTION

     "rms_cur" computes the RMS difference between two atom
     selections without performing any fitting.

  USAGE

     rms_cur (selection), (selection)

  SEE ALSO

     fit, rms, intra_fit, intra_rms, intra_rms_cur, pair_fit

Unfortunately it leaves out the bit included in:
  PyMOL>help fit

  DESCRIPTION

     "fit" superimposes the model in the first selection on to the model
     in the second selection.  Only matching atoms in both selections
     will be used for the fit.

This is still not quite as clear as it could be. "Matching atoms" are
those that are identical in chain, residue name, residue number, atom
name, segid and alternate position identifier (if present). So although
you had a two lists of 35 C-alphas, none of them matched in all of those
aspects.

At the end of March, I had announced on this mailing list a script that
takes care of this problem when trying to use the "fit" command and it
was easy to modify it to use the "rms_cur" command. You can download
them from my pymol script site:

  http://adelie.biochem.queensu.ca/~rlc/work/pymol

the scripts are called fitting.py and rms_current.py (I couldn't think
of a better name quickly). The usage is:

  rms_current object1, selection1, object2, selection2

For your case:

  rms_current 1gh2, n. ca & i. 65-99, 1kao, n. ca & i. 64-98

Note that the object names must not appear in the selections because
the atom specifications from object1 are (temporarily) changed to match
object2 and selection2. I know this is sort of ugly, but it was to avoid
using the slow "alter" command on both selections. 

I hope this helps.

Cheers,
Rob
-- 
Robert L. Campbell, Ph.D.                         <rl...@po...>
Senior Research Associate                            phone: 613-533-6821
Dept. of Biochemistry, Queen's University,             fax: 613-533-2497
Kingston, ON K7L 3N6  Canada       http://adelie.biochem.queensu.ca/~rlc
    PGP Fingerprint: 9B49 3D3F A489 05DC B35C  8E33 F238 A8F5 F635 C0E2