Kristina -
Thanks much for the reminder. I pulled up our correspondence in my email,
and I indeed do have a test case from you. As it turned out, looking at the
email, this bug in pmemd is not absolutely critical because there is a
workaround; all you have to do is omit the DISANG line and put the contents
of the DISANG file in a &rst namelist. I have not heard back from Cristoph,
so have not proceeded with work on a fix. If the workaround is unsuitable
for anyone, please let me know, but Kristina basically worked out a way to
have nmropt weights (&wt namelist), restraints (&rst namelist), and "ntr"
coordinate constraints (ntr=1, group input following all the namelists)
simultaneously. I will issue a patch if indeed someone cannot get their
work done on account of all this. Otherwise, I will be trying to track
sander input behaviour in the next release; there are some new
simplifications in amber 8 that I don't yet allow in pmemd (documented) that
I should provide; I don't know what other input changes are coming at this
point in time. It would be nice if we had less ambiguity about the various
combinations of input; trouble is that it is also nice to not break existing
inputs.
Regards - Bob
----- Original Message -----
From: "Furse, Kristina Elisabet" <kristina.e.furse.vanderbilt.edu>
To: <amber.scripps.edu>
Sent: Friday, June 17, 2005 3:16 PM
Subject: Re: AMBER: nmropt =1 AND ntr=1 in pmemd
> Hi Robert-
>
> We have indeed been down this road before--I think I sent you a test case.
> Here's our thread from the mailing list archive if that helps:
>
> http://amber.ch.ic.ac.uk/archive/200402/0066.html
>
> I'll search my email to see if I can find the resolution, since I can't
> recall
> what happened!
>
> Kristina
>
> Quoting Robert Duke <rduke.email.unc.edu>:
>
>> Cristoph -
>> Which version of pmemd? Have you tried the inputs with sander 7 or
>> sander
>> 8, just a short run to see if it works? If you would like me to take a
>> more
>>
>> detailed look, please put the prmtop/inpcrd/mdin and additional
>> redirections
>>
>> in a tarball and mail it directly to me (at rduke.email.unc.edu) and I
>> will
>> look at it, and if there is a deficiency in pmemd, patch it and send you
>> the
>>
>> patch. The story on these more complicated input modes is that starting
>> from the first pmemd releases, I made the namelist reading code more
>> reliable by rewinding after every namelist read, thus making namelist
>> reading position independent. Now, I know that these two types of
>> restraints work alone, but I am not sure whether I have ever been asked
>> if
>> they work together. Anyway, I will be glad to look at it, but because we
>> don't have this sort of test case, it would be helpful if you send your
>> input.
>> Regards - Bob Duke
>>
>> ----- Original Message -----
>> From: "Christoph Brockmann" <brockmann.fmp-berlin.de>
>> To: <amber.scripps.edu>
>> Sent: Tuesday, June 14, 2005 5:33 AM
>> Subject: AMBER: nmropt =1 AND ntr=1 in pmemd
>>
>>
>> > Dear mailing list,
>> >
>> > I try to run amber8 pmemd with both nmr AND positional restraints.
>> > However, it seems impossible to get pmemd to read both types of
>> > restraints.
>> >
>> > If I give the group information after the redirects (in a sander7
>> > fashion)
>> >
>> > %<
>> >
>> >
>> > LISTIN=SOLVRUN1_c.restraints
>> >
>> >
>> >
>> > LISTOUT=SOLVRUN1_c.violations
>> >
>> >
>> > DISANG=test.rst
>> > Positional
>> > restraints
>> >
>> > 100
>> >
>> >
>> >
>> > FIND
>> >
>> >
>> > SD * *
>> > *
>> >
>> > SEARCH
>> >
>> >
>> > RES 1
>> > 10
>> >
>> > END
>> >
>> >
>> >
>> > END
>> >
>> >
>> >
>> > STOP
>> >
>> >
>> > %<
>> >
>> > I get:
>> >
>> > %<
>> > 5. REFERENCE ATOM COORDINATES
>> >
>> >
>> >
>> >
>> > ----- READING GROUP 1; TITLE:
>> > LISTIN=SOLVRUN1_c.restraints
>> >
>> >
>> > rfree: Error decoding variable 1 3 from:
>> > LISTOUT=SOLVRUN1_c.violations
>> > %<
>> >
>> > If I give the Group BEFORE the redirects the group information is read
>> > and understood but the redirects themselves are not found anymore
>> > resulting in absolutely NO nmr-restraints. Is there a way to solve this
>> > problem?
>> >
>> > Christoph
>> >
>> >
>> >
>> > -----------------------------------------------------------------------
>> > The AMBER Mail Reflector
>> > To post, send mail to amber.scripps.edu
>> > To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
>> >
>>
>>
>> -----------------------------------------------------------------------
>> The AMBER Mail Reflector
>> To post, send mail to amber.scripps.edu
>> To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
>>
>
>
> -----------------------------------------------------------------
> Kristina E. Furse
> Department of Chemistry
> Center for Structural Biology
> Vanderbilt University
> Email: kristina.e.furse.Vanderbilt.Edu
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber.scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Fri Jun 17 2005 - 23:53:00 PDT
