From chemistry-request*- at -*ccl.net Fri Jul  2 10:59:30 2004
Received: from mhub-w3.tc.umn.edu (mhub-w3.tc.umn.edu [160.94.160.33])
	by server.ccl.net (8.12.8/8.12.8) with ESMTP id i62FxTFY003303
	for <chemistry<<at>>ccl.net>; Fri, 2 Jul 2004 10:59:30 -0500
Received: from pobox.com (abacavir.pharmacy.umn.edu [160.94.172.208] (may be forged)) by mhub-w3.tc.umn.edu with ESMTP; Fri, 2 Jul 2004 11:04:53 -0500 (CDT)
X-Umn-Remote-Mta: [N] abacavir.pharmacy.umn.edu [160.94.172.208] #+HF+LO
Message-ID: <40E588A8.7080507<<at>>pobox.com>
Date: Fri, 02 Jul 2004 11:09:12 -0500
From: Eric Bennett <ericb<<at>>pobox.com>
User-Agent: Mozilla/5.0 (X11; U; IRIX64 IP30; en-US; rv:1.4.1) Gecko/20040326
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: Michel Petitjean <ptitjean<<at>>itodys.jussieu.fr>, chemistry<<at>>ccl.net
Subject: Re: software to convert  XYZ format to PDB ?
References: <200407020730.i627U68E052048<<at>>ds10.itodys.jussieu.fr>
In-Reply-To: <200407020730.i627U68E052048<<at>>ds10.itodys.jussieu.fr>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Spam-Status: No, hits=0.0 required=7.5 tests=none autolearn=no version=2.61
X-Spam-Checker-Version: SpamAssassin 2.61 (1.212.2.1-2003-12-09-exp) on 
	servernd.ccl.net

Michel Petitjean wrote:

>To: chemistry<<at>>ccl.net
>Subject: CCL:software to convert  XYZ format to PDB ?
>
>Konrad Hinsen <hinsen<<at>>cnrs-orleans.fr> wrote:
>  
>
>> If I had to do this (fortunately I don't), I'd first identify the
>>
>>peptide planes as patterns of pairwise distances (they are quite 
>>rigid). I'd then identify the sidechains atom by atom starting from the 
>>C-alphas, again by distance criteria. Not a trivial task, but not an 
>>impossible one either.
>>    
>>
>
>Sure... if you have the side chains ! (were not indicated in the example)
>It needs other assumptions: true residues, no HETATM, lexicographical
>ordering coherent with the PDB conventions...
>  
>
There are programs (such as ARP/wARP) which start with NO atom information, not even coordinates, and can correctly build a protein structure based on an electron density map using pattern recognition techniques.  The ambiguities there, such as ASP vs. ASN, are not present when you know the elements, as we do here.  ARP/wARP is also gaining ligand building capabilities and already handles water molecules.

ARP is part of CCP4 so you get source code with it.
http://www.embl-hamburg.de/ARP/

Of course if you have any unexplained data (whether electron density or known atom positions) that do not fit the pattern of an amino acid, you can just leave them alone and inform the user.



-- 
Eric Bennett, Center for Drug Design, U of Minnesota