From vijay@rose.chem.wesleyan.edu  Fri Jul  1 12:04:48 1994
Received: from rose.chem.wesleyan.edu  for vijay@rose.chem.wesleyan.edu
	by www.ccl.net (8.6.9/930601.1506) id LAA10321; Fri, 1 Jul 1994 11:26:11 -0400
Received: from jellyfish.chem.wesleyan.edu by rose.chem.wesleyan.edu (AIX 3.2/UCB 5.64/4.03)
          id AA20763; Fri, 1 Jul 1994 11:24:17 -0400
Date: Fri, 1 Jul 1994 11:24:17 -0400
Message-Id: <9407011524.AA20763@rose.chem.wesleyan.edu>
X-Sender: vijay@rose
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
To: chemistry@ccl.net
From: vijay@rose.chem.wesleyan.edu (S. Vijayakumar)
Subject: RMSD
X-Mailer: <PC Eudora Version 1.4b17>


mikes@bioch.ox.ac.uk wrote:

: I am looking for a software package which will enable me to
: least-squares-fit and find the rmsds between two structures.  The
: complicating factor is that I want to do it on structures of different
: amino acid sequences, to fit on the backbone and all relevent atoms, and
: give an rmsd based on this.
: 
: I have tried XPLOR but it seems to insist the molecules are
: identical.  I would also like to have the backbone rmsd printed out on a
: residue-by-residue basis.  
:
----Other stuff deleted----------------

and  brian@bert.chem.wsu.edu responded,
	
>                        I've run into the same problem with the proteins
>	I work with. Since you seem to be primarily interested in
>	the backbone RMSD (as I was), then why don't you edit out
>	all the side chains (i.e. make all the side chains =GLY).

        The origin of the problem is as follows:

        Two proteins of length 'n' and 'm' can be superposed in n-m+1 ways, 
assuming connectivity is retained and the superposition is unidirectional, 
i.e. N-terminal end to C-terminal end. The majority of the programs 
circumvent this problem by requiring that the molecules be identical or 
contain the same number of atoms! Thus if you can specify the residues 
ranges for each protein
and apply a suitable selection criteria such that there is a one-to-one 
correspondence between the two sets of atoms (this necessitates equal number 
of atoms), then most commercial programs can superpose the two structures. I 
am, however, not aware if they can spit out residue based RMS.

        If you can write a script to extract only the necessary coords., I 
have a home grown program that can superpose any two structures and spit out 
residue based RMS. Unfortunately it does not read a variety of input formats 
as is, but you can modify the code to read in your coords. with minimal 
effort.  If you send me e-mail I can mail you the code. 

        Hope this helps.

Best wishes,

Vijay
S. Vijayakumar
Department of Chemistry
Wesleyan University
Middletown, CT 06459
E-mail: vijay@rose.chem.wesleyan.edu
Voice: (203) 347-9411, x3196
Fax: (203) 344-7960