VMD-L Mailing List

From: René Hafner TUK (hamburge_at_physik.uni-kl.de)
Date: Fri Oct 23 2020 - 04:15:15 CDT

Sure just send me any Cloud link and I can have a look at the pictures
(or via email only to me if  the image sizes are not too big.)

I don't know about your available computing resources but from my
experiences a Cluster generally has a few nodes with 1TB of RAM which
would be enough. I'd try it there.

Give the qwrap tool a try! :)

On 10/23/2020 10:33 AM, McGuire, Kelly wrote:
> Thanks for the response! Each DCD is about 10 GB. So, 500 GB for the
> whole trajectory.
>
> Haven't ever tried qwrap. I thought it was mostly for orthorhombic
> water box shapes.  Maybe it's worth a try.
>
> I couldn't attach the pictures to the first email.  Is there a way I
> can send them to you for better context of the description in my
> email? It might clarify the problem a bit more.  2-3 of the segments
> are hanging outside of the water box on one side and the 4th segment
> is hanging outside the water box on the other side.  Can't seem to get
> all segments back together in the center of the box no matter what set
> of commands I use....
>
> /Dr. Kelly L. McGuire/
>
> /PhD //Biophysics/
>
> /Department of Physiology and Developmental Biology/
>
> /Brigham Young University/
>
> /LSB 3050/
>
> /Provo, UT 84602/
>
>
> ------------------------------------------------------------------------
> *From:* René Hafner TUK <hamburge_at_physik.uni-kl.de>
> *Sent:* Friday, October 23, 2020 2:25 AM
> *To:* McGuire, Kelly <mcg05004_at_byui.edu>; VMD Mailing LIst
> <vmd-l_at_ks.uiuc.edu>
> *Subject:* Re: vmd-l: PBC Wrapping
>
> Dear Kelly,
>
>
> First recommendation:
>     Use Jérôme Hénin qwrap tool for wrapping/unwrapping of trajectories.
>
> https://github.com/jhenin/qwrap <https://github.com/jhenin/qwrap>
>
>
> It is a lot faster than the pbctools wrap command.  (:
>
>
>
> 2. What file sizes are we talking about here? Dozens of GB or TB?
>
> From my point of view:
>
>     If you have a PC or cluster node with enough RAM it might be
> better to load the whole 2µs DCD at once.
>
>
> I'd say if the simulation ran properly regarding the Box size (NVT I
> guess) there shouldn't be too much of a problem regarding wrapping the
> trajectory. It might just be that the dynamics are unreasonable if the
> protein segment interact with other parts of themselves by PBC and
> therefore may invalidate the results.
>
>
> Kind regards
>
> René
>
>
>
> On 10/22/2020 11:06 PM, McGuire, Kelly wrote:
>> I know many people have asked a similar question and the topic is
>> quite exhausted. However, I have tried multiple combinations of PBC
>> wrap that I have found int he mailing list or in the manual and they
>> aren't working this case.
>>
>> Problem Background:
>>
>> I just started a postdoc and have been asked to help one of the grad
>> students to get the protein joined back together and re-centered in
>> the water box.  He has run a 2-microsecond simulation of a multimeric
>> protein (4 segments) in a water box. Two problems I noticed
>> immediately is that his water box is too small for the protein (in my
>> opinion) and second, he did not have wrap turned on in his config
>> file.  He has chopped this 2-microsecond simulation into
>> 40-nanosecond long DCD files, so he has 50 DCD files that need to be
>> wrapped.  In some of the DCD files I can get the protein joined back
>> together and re-centered, but there are many that I cannot. I have
>> pictures of one such very difficult case (email won't go through with
>> attachments, how to send pics?).  Blue is segname PROA, Red is
>> segname PROB, Orange is segname PROC, and Green is segname PROD. 
>> These pictures represent ~1.8-microseconds into the simulation. 
>> Pic-1 and Pic-2 are the first frame (front and top-down views
>> respectively). Pic-3 is the top-down view of the last frame.  All of
>> the protein segments are constantly close to the boundary conditions.
>> I can only get 2-3 of the segments to join back up before centering
>> the water box. There is always one of the segments that won't join
>> back with the rest of the protein.
>>
>> Questions:
>>
>> 1) Is it possible to do a wrap on the whole 2-microseconds DCD files
>> at once? Meaning load all of the DCD files and wrap the protein? Or
>> is it better to do each DCD file individually?
>>
>> 2) Is this an impossible case to wrap because the water box is too small?
>>
>> 3) Any specific suggestions on how to get this protein put back
>> together and re-centered?
>>
>> 4) In the first few DCD files, I used:
>>
>> pbc wrap -center com -centersel "protein and segname PROB" -compound
>> residue -all
>>
>> pbc wrap -center com -centersel "protein and segname PROD" -compound
>> residue -all
>>
>> pbc wrap -center com -centersel "protein" -compound residue -all
>>
>>
>> This worked just fine. Why would this not work for all 50 DCD files
>> for the whole 2 microseconds? Are different wrapping commands needed
>> for each DCD of the same long simulation?
>>
>>
>> /Dr. Kelly L. McGuire/
>>
>> /PhD //Biophysics/
>>
>> /Department of Physiology and Developmental Biology/
>>
>> /Brigham Young University/
>>
>> /LSB 3050/
>>
>> /Provo, UT 84602/
>>
>>
> --
> --
> Dipl.-Phys. René Hafner
> TU Kaiserslautern
> Germany 

-- 
--
Dipl.-Phys. René Hafner
TU Kaiserslautern
Germany