different krama for different parts of the protein?

Hi all,

I'm trying to refine a structure where a portion of the protein does not
have NOESY restraints associated with it.  However, I would like to apply
the database potentials so that the structures that I generate at least
have reasonable ramachandran plots.  However, if I use a standard krama of
1, it seems like my structures still have poor ramachandran statistics.
If I increase krama to 20, I get better ramachandran stats, but I think
that the number of violated NOE restraints have increased.  Is there a way
to restrain a portion of a protein with a stronger database potential
while the rest of the protein is restrained with a weaker potential?
I'm using a script modified from the refinement protocol in
eginput/gb1_rdc.

Thanks,
-Greg Zornetzer
gaz <at> nmrfam.wisc.edu

Re: different krama for different parts of the protein?

Hi Greg,

Unfortunately, the answer is no.  Surprisingly, you're the first  
person to ever
ask for this feature.  I'll look into it and see what the most  
elegant way of
implementing it would be.

But regarding the problem you're running into with NOEs, I can offer  
some
more useful advise.  Cases where NOE violations appear when the rama
term is turned on are almost always the result of a bad NOE  
restraint.  So
take a close look at the violations that appear when you refine with  
krama = 20,
and you will probably find an error.

Let me know if you're still having problems.

--JK

On Jan 6, 2006, at 3:13 PM, Gregory Zornetzer wrote:

> Hi all,
>
> I'm trying to refine a structure where a portion of the protein  
> does not
> have NOESY restraints associated with it.  However, I would like to  
> apply
> the database potentials so that the structures that I generate at  

how to incorporate into MTSL into a protein?

Re: how to incorporate into MTSL into a protein?]

On Mon, 2006-01-30 at 16:04 -0600, Wu, Bainan wrote:
> Hi, all
> 
>  
> 
> I am a newer of xplor-nih, so I hope this is not a stupid question. I
> am trying to incorporate a MTSL into the structure of my protein.
> 
Adding ligands into a structure was tricky for me.  But it is doable.  

> Luckily, I found xplor-nih had the topology file of the MTSL and also
> included a protocol in PSF_generation folder of xplor-nih-2.12 talking
> about how to incorporate it into protein. I test the demo, run
> addAtoms.py and it looks work pretty good. However, after I change the
> script with the sequence and pdb file of my protein, it shows the
> warning of “atom not found in structure” for every atom in pdb file.
> It seems that PDBTool in the 
> 
> Python file does not read my pdb file at all. Can anybody tell me why?
> Is that because of the format of pdb file? If yes, could you give me
> some hints about how to fix it? 
> 
One problem that I ran into had to do with segment identifiers.  If the
seg ID's in your input PDB don't match the chain ID's in the PSF file,
then XPLOR-NIH won't find the atoms. Does your PDB file have segment
ID's in it? (They're in something like rows 70-74) - past the x, y, z,
occupancy, and B-factor columns.  I notice that the example in
eginput/PSF_generation uses no segment ID.  If you do need a segment ID,
there's an option to seqToPSF() called segName that should help. See 

 http://nmr.cit.nih.gov/xplor-nih/doc/current/python/ref/psfGen.html

I hope this helps.

-Greg Zornetzer
gaz <at> nmrfam.wisc.edu

--

-- 

RE: how to incorporate into MTSL into a protein?

Thanks for everybody, especially for Greg. 

I go back to check my pdb file and then find possible mistakes.
The initial structure uses two long linkers to connect three chains.
When I prepare my pdb file, I delete these linkers without giving the
different chain number. This may be the point. 

Here, I have another question. If there is a jump on residue number due
to the chain break, can addAtom.py read through sequence file to match
the corresponding atoms in pdb file?

Best Regards!  

bwu

-----Original Message-----
From: Greg Zornetzer [mailto:gaz <at> nmrfam.wisc.edu] 
Sent: Monday, January 30, 2006 4:29 PM
To: Wu, Bainan
Subject: Re: [Xplor-nih] how to incorporate into MTSL into a protein?

On Mon, 2006-01-30 at 16:04 -0600, Wu, Bainan wrote:
> Hi, all
> 
>  
> 
> I am a newer of xplor-nih, so I hope this is not a stupid question. I
> am trying to incorporate a MTSL into the structure of my protein.
> 
Adding ligands into a structure was tricky for me.  But it is doable.  

> Luckily, I found xplor-nih had the topology file of the MTSL and also
> included a protocol in PSF_generation folder of xplor-nih-2.12 talking
> about how to incorporate it into protein. I test the demo, run
> addAtoms.py and it looks work pretty good. However, after I change the
> script with the sequence and pdb file of my protein, it shows the
> warning of "atom not found in structure" for every atom in pdb file.
> It seems that PDBTool in the 
> 
> Python file does not read my pdb file at all. Can anybody tell me why?
> Is that because of the format of pdb file? If yes, could you give me
> some hints about how to fix it? 
> 
One problem that I ran into had to do with segment identifiers.  If the
seg ID's in your input PDB don't match the chain ID's in the PSF file,
then XPLOR-NIH won't find the atoms. Does your PDB file have segment
ID's in it? (They're in something like rows 70-74) - past the x, y, z,
occupancy, and B-factor columns.  I notice that the example in
eginput/PSF_generation uses no segment ID.  If you do need a segment ID,
there's an option to seqToPSF() called segName that should help. See 

 http://nmr.cit.nih.gov/xplor-nih/doc/current/python/ref/psfGen.html

I hope this helps.

-Greg Zornetzer
gaz <at> nmrfam.wisc.edu

Re: how to incorporate into MTSL into a protein?

Hi--

> 
> I go back to check my pdb file and then find possible mistakes.
> The initial structure uses two long linkers to connect three chains.
> When I prepare my pdb file, I delete these linkers without giving the
> different chain number. This may be the point. 
> 

sounds like you're making progress...

> Here, I have another question. If there is a jump on residue number due
> to the chain break, can addAtom.py read through sequence file to match
> the corresponding atoms in pdb file?
> 

For this, modify the addAtoms.py script as follows:

1) split the seq string into two, named say seq1 and seq2, broken at the
   chain break. 
2) call psfGen.seqToPSF twice: first using seq1 and startResid=1
                               second using seq2 and the appropriate startResid

hope this helps--

Charles

how to incorporate into MTSL into a protein? . .

Hi, all,

Now, I have made some progress. According to charles' suggestion, I
split the seq string into two parts and then start PDBTool after two
steps of import psfGen with appropriate startResid. It looks right and I
finally get the output of pdb and psf. However, I notice there are some
error reports which happen during the script running. I list some of
them as following:

"PARRDR-ERROR: duplication of nonbonded entry CS2", or, 
"PARRDR-ERROR: duplication of bond entry C3S8 HA", or, 
"PARRDR-ERROR: duplication of angle C3S6 CSN2 C3S7"

I check the output pdb file and find that the rotatable bonds between
cysteine and the ring of MTSL are missing in the final structure. Is
this related to the above error reports? Maybe not, because it seems
these bonds are missing as well in the demo output. 

Since I don't want to go back again to correct my mistakes, here I have
another question, that is "whether is this structure without the bond
linking between cys and MTSL useful for further ensemble calculation or
back calculation?"

Thanks for everybody in advance!

bwu

-----Original Message-----
From: 
Sent: None
To: Wu, Bainan
Subject: Re: [Xplor-nih] how to incorporate into MTSL into a protein? .
.

In-reply-to:
<5CB39BCA5724F349BCB748675C6CA1A20381957F <at> SJMEMXMB02.stjude.sjcrh.local>
References:
<5CB39BCA5724F349BCB748675C6CA1A20381957F <at> SJMEMXMB02.stjude.sjcrh.local>
Comments: In-reply-to "Wu, Bainan" <Bainan.Wu <at> stjude.org>
	message dated "Mon, 30 Jan 2006 22:49:00 -0600."
X-Mailer: MH-E 7.85; nmh 1.1; GNU Emacs 21.4.1
From: Charles <at> schwieters.org
Date: Tue, 31 Jan 2006 09:48:12 -0500
Message-Id: <E1F3wnk-0001Pp-UP <at> beige.cit.nih.gov>
Cc: xplor-nih <at> nmr.cit.nih.gov
X-BeenThere: xplor-nih <at> nmr.cit.nih.gov
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: xplor-NIH help list <xplor-nih.nmr.cit.nih.gov>
List-Unsubscribe: <http://dcb.cit.nih.gov/mailman/listinfo/xplor-nih>,
	<mailto:xplor-nih-request <at> nmr.cit.nih.gov?subject=unsubscribe>
List-Archive: <http://cake.cit.nih.gov/pipermail/xplor-nih>
List-Post: <mailto:xplor-nih <at> nmr.cit.nih.gov>
List-Help: <mailto:xplor-nih-request <at> nmr.cit.nih.gov?subject=help>
List-Subscribe: <http://dcb.cit.nih.gov/mailman/listinfo/xplor-nih>,
	<mailto:xplor-nih-request <at> nmr.cit.nih.gov?subject=subscribe>
Sender: xplor-nih-bounces <at> nmr.cit.nih.gov
Errors-To: xplor-nih-bounces <at> nmr.cit.nih.gov
Return-Path: xplor-nih-bounces <at> nmr.cit.nih.gov
X-OriginalArrivalTime: 31 Jan 2006 14:48:27.0800 (UTC)
FILETIME=[653A8180:01C62675]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi--

> 
> I go back to check my pdb file and then find possible mistakes.
> The initial structure uses two long linkers to connect three chains.
> When I prepare my pdb file, I delete these linkers without giving the
> different chain number. This may be the point. 
> 

sounds like you're making progress...

> Here, I have another question. If there is a jump on residue number
due
> to the chain break, can addAtom.py read through sequence file to match
> the corresponding atoms in pdb file?
> 

For this, modify the addAtoms.py script as follows:

1) split the seq string into two, named say seq1 and seq2, broken at the
   chain break. 
2) call psfGen.seqToPSF twice: first using seq1 and startResid=1
                               second using seq2 and the appropriate
startResid

hope this helps--

Charles
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.8
<http://mailcrypt.sourceforge.net/>

iD8DBQFD33isPK2zrJwS/lYRArd9AKCBihN3JIDgAHD9eklexpguOtn8sACdEo5t
1fITcR12jVfZfEmhrd2JPhg=
=Qc98
-----END PGP SIGNATURE-----