headers
Wenyong Tong | 2 Jan 2007 01:04
Picon
Favicon

Re: for Marvin


Hi 

I am going to run Marvin again. Thanks.

I have 2-D NOESY and ROESY of small cyclic peptides,
but I use sparky for assignment. 

Happy New Year to every XPLOR-NIH users.

wy

--- John Kuszewski <johnk <at> mail.nih.gov> wrote:

> 
> On Dec 23, 2006, at 6:16 PM, Wenyong Tong wrote:
> 
> >
> > Hi
> >
> > I ran the runMarvin for CVN, but I did not get the
> > cvn_xplor_noes.tbl finally. Some of messages are
> as
> > follows:
> > TCL>readShiftAssignments \
> > TCL>    -fileName "cvn_3dc_final.shiftAssignments"
> \
> > TCL>    -pot $noe
> > Error opening input file
> > cvn_3dc_final.shiftAssignments
(Continue reading)

Permalink | Reply | 0 comments |
headers
Barbara Richter | 23 Jan 2007 13:22
Picon
Picon
Favicon

rdcs, shake

Hi,

Two questions...

When restraining RDCs on the sidechain, which types do you use? Is it ok to 
use the "CACO" type for carbon-carbon RDCs on the sidechain? And the "CAHA" 
type for carbon-hydrogen RDCs? I guess the question I am trying to ask is 
whether it is ok to assume that the bond lengths are the same in these 
cases.

Also, what is the easiest way to keep bond lengths (but not angles) 
constant. Is there an equivalent of the SHAKE statement in XPlor-NIH?

Thanks!
Barbara
Permalink | Reply | 1 comment |
headers
Charles | 23 Jan 2007 17:34

Re: rdcs, shake


Hello Barbara--

> 
> When restraining RDCs on the sidechain, which types do you use? Is it ok to 
> use the "CACO" type for carbon-carbon RDCs on the sidechain? And the "CAHA" 
> type for carbon-hydrogen RDCs? I guess the question I am trying to ask is 
> whether it is ok to assume that the bond lengths are the same in these 
> cases.

For CH RDCs, the answer is probably: yes, it's safe to assume CAHA type.
For CC, probably not- you probably should use a bond length of 1.38 for
those. Let me know if you'd like me to help incorporate that value into
rdcPotTools

> 
> Also, what is the easiest way to keep bond lengths (but not angles) 
> constant. Is there an equivalent of the SHAKE statement in XPlor-NIH?
> 

You could use the "bendtorsion" hingetype within the IVM- this allows
both torsion and angle motion. What are you trying to do with this?

best regards--
Charles
Permalink | Reply | 0 comments |
headers
R.M. van der Werf | 25 Jan 2007 14:34
Picon

problems xplor-NIH on clustercomputer

Hi,

I am trying to run xplor-NIH (version 2.16.0) on a cluster computer, but
having problems.
Here's what I did do:

1) I installed xplor-NIH in my own user directory, and manipulated the
PBS script mentioned at the xplor-NIH site to match the cluster here.

In the script I specified the number of cpus to be used #PBS -l ncpus=##

2) When I submit the job:
qsub myjob.pbs

The task starts running (confirmed by mail to my account)
Then the timer locks at the value of 00:00:00 but somehow indicates a
running status, output of qstat:

Job id              Name             User             Time Use S Queue
------------------- ---------------- ---------------- -------- - -----
14474.beoclus       RDC_short        ramon            00:20:57 R
dque           
14477.beoclus       RDC_short_NIH    ramon            00:00:00 R
dque           

First job is the one I am running with the 'old' xplor (working
perfectly, but can only use one cpu per script)
The last one is with xplor-NIH after 10 minutes of running.

On the Node that is running two processes named xplor are spawned by
(Continue reading)

Permalink | Reply | 1 comment |
headers
Charles | 25 Jan 2007 16:03

Re: problems xplor-NIH on clustercomputer


Hello Ramon--

> 
> On the Node that is running two processes named xplor are spawned by
> xplor-NIH. Both processes generate output PDBs, whereas I only want one.
> (Just like I would like to run ONE xplor-NIH on this node)
> 

If, in your PBS script, you have a line like this 
cat $PBS_NODEFILE $PBS_NODEFILE > $XPLOR_MACHINES
change it to
cat $PBS_NODEFILE > $XPLOR_MACHINES

> The logfile remains empty. Probably because writing is hampered by two
> processes trying to write to this file.

what is in the file script.out.0? This is the Xplor-NIH logfile as
specified in the PBS script.

> 
> When I type qdel 14477 (delete the xplor-NIH process)
> The process is eliminated from the queue, but both processes keep on
> running on the node.

This is a problem for your PSB administrator- it should clean up after
rogue jobs. That said, you could simplify your ``run xplor'' statement
for your case to make this more bulletproof. Change it to

 $XPLOR $PY -parallel -o $LOGFILE $XPLOR_SCRIPT
(Continue reading)

Permalink | Reply | 0 comments |
headers
Tomasz Cierpicki | 26 Jan 2007 21:31

weights for energy terms

Hi

I have question about setting weights for experimental energy terms.
Unexpectedly, regardless of using different values for knoe or kcdi I am
getting exactly the same structures.

Here is example of my input file:

evaluate ($knoe  = 30.0)   {* noes potential         *}

noe
   ceiling 1000
   averaging  all sum
   potential  all square
   scale      all $knoe
   sqconstant all 1.0
   sqexponent all 2
end

And this is part of my output file:

X-PLOR>evaluate ($knoe  = 30.0)   {* noes potential         *}
EVALUATE: symbol $KNOE set to   30.00000     (real)

...

X-PLOR>noe
NOE>  ceiling 1000
NOE>  averaging  all sum
NOE>  potential  all square
(Continue reading)

Permalink | Reply | 1 comment |
headers
John Kuszewski | 26 Jan 2007 21:45
Picon

Re: weights for energy terms

Most likely, your script is setting kNOE in more than one place.  If  
you send us your script, we
could find the problem.

--JK

On Jan 26, 2007, at 3:31 PM, Tomasz Cierpicki wrote:

> Hi
>
> I have question about setting weights for experimental energy terms.
> Unexpectedly, regardless of using different values for knoe or kcdi  
> I am
> getting exactly the same structures.
Permalink | Reply | 0 comments |

Gmane