Re: [AMBER] cpptraj - trouble with autoimage and triclinic truncated octahedron DCD from NAMD

From: Daniel Roe <daniel.r.roe.gmail.com>
Date: Wed, 30 Nov 2016 14:18:17 -0500

Hi,

Please try the GitHub version of cpptraj:
https://github.com/Amber-MD/cpptraj. I just added two new keywords to
'trajin' for CHARMM trajectories: 'ucell' to force reading box info as
unit cell params, and 'shape' to force reading box info as shape
matrix. So you should be able to do:

trajin mytraj.dcd ucell

to get it to read the box as unit cell params. If you get a chance to
try it and it works (or doesn't), let me know. Thanks,

-Dan

On Tue, Nov 29, 2016 at 4:02 PM, Daniel Roe <daniel.r.roe.gmail.com> wrote:
> If you can use the older version of cpptraj for now that is the
> quickest workaround. I will try to get something going over the next
> day or so. Thanks,
>
> -Dan
>
> On Tue, Nov 29, 2016 at 2:34 PM, Brian Radak
> <brian.radak.accts.gmail.com> wrote:
>> Aha! I am definitely not using a symmetric shape matrix (which is weird,
>> because I got the coordinates from charmm-gui).
>> I should have posted the XSC contents earlier:
>>
>> # NAMD extended system configuration output file
>> #$LABELS step a_x a_y a_z b_x b_y b_z c_x c_y c_z o_x o_y o_z
>> 0 83.5014512836 0 0 -27.8338170946 78.7259232555 0 27.8338170946
>> 39.3629616277 68.1786494757 0 0 0
>>
>> The length parameter is definitely that first entry, but the others are
>> scaled values. I had found a page describing this earlier via google, but
>> have since lost it.
>>
>> Is there a fix for this? Can I somehow give parmbox a symmetric matrix for
>> the same shape?
>>
>>
>> On Tue, Nov 29, 2016 at 1:26 PM, Daniel Roe <daniel.r.roe.gmail.com> wrote:
>>
>>> Hi,
>>>
>>> Hmm...I'm not sure how NAMD stores unit cell info in the DCD
>>> trajectory. According to CHARMM, since version 22 the symmetric shape
>>> matrix is stored instead of unit cell parameters. The weird angle
>>> you're getting for truncated octahedron makes me suspect that the unit
>>> cell information in your trajectory is actually unit cell parameters.
>>> This can be easily tested if you have a version of cpptraj laying
>>> around < 16.16 - prior to this cpptraj didn't understand the shape
>>> matrix info output by CHARMM. If this is indeed the case I'll have to
>>> update cpptraj so it can figure out (or be told) what kind of unit
>>> cell parameters are in a DCD file.
>>>
>>> -Dan
>>>
>>> On Tue, Nov 29, 2016 at 1:56 PM, Brian Radak
>>> <brian.radak.accts.gmail.com> wrote:
>>> > I just downloaded AmberTools16 today, Cpptraj V16.16.
>>> >
>>> > There are probably a number of places weird things could happen. Maybe
>>> the
>>> > following observations are relevant:
>>> >
>>> > 1) the CHARMM PSF format does not include *any* box information
>>> > 2) the extended system (xsc) file that I used in my simulation weirdly
>>> has
>>> > all zeros for the angles (I guess because NAMD uses (can use?) a full
>>> > vector specification instead?)
>>> > 3) I have little idea as to how the DCD trajectories store box
>>> information
>>> > 4) this was a constant volume simulation
>>> >
>>> > On Tue, Nov 29, 2016 at 12:48 PM, Daniel Roe <daniel.r.roe.gmail.com>
>>> wrote:
>>> >
>>> >> Hi,
>>> >>
>>> >> What version of cpptraj are you using? Something is clearly off since
>>> >> those are definitely not angles for a truncated octahedron cell.
>>> >>
>>> >> -Dan
>>> >>
>>> >> On Tue, Nov 29, 2016 at 1:24 PM, Brian Radak
>>> >> <brian.radak.accts.gmail.com> wrote:
>>> >> > I have a NAMD DCD trajectory and PSF generated with triclinic boundary
>>> >> > conditions corresponding to a truncated octahedron, but am having
>>> trouble
>>> >> > getting the trajectory into "familiar" form via cpptraj (which is
>>> >> > apparently the only utility around that can handle this procedure?).
>>> >> >
>>> >> > I did the following:
>>> >> >>> cpptraj -p my.psf -y my.dcd
>>> >> >
>>> >> > Which gave the following helpful warning:
>>> >> >
>>> >> > Warning: Trajectory box type is 'Non-orthogonal' but topology box
>>> type is
>>> >> > 'None'.
>>> >> > Warning: Setting topology box information from trajectory.
>>> >> >
>>> >> > Then
>>> >> >>> autoimage
>>> >> >>> trajout my.rewrap.dcd
>>> >> >>> go
>>> >> >
>>> >> > However, loading the output dcd into VMD instead shows an orthorhombic
>>> >> cell
>>> >> > with missing slabs at a diagonal. I checked that the
>>> >> a/b/c/alpha/beta/gamma
>>> >> > were what I thought they should be:
>>> >> >>> parmbox
>>> >> > Box: 'Non-orthogonal' XYZ= { 83.503 83.503 83.503 } ABG= {
>>> >> > 89.54 89.54 90.46 }
>>> >> >
>>> >> > Which did not seem right, so I tried again with:
>>> >> >>> parmbox truncoct 83.503
>>> >> >
>>> >> > but this has the same result. I saw indications in the list archive
>>> that
>>> >> > similar things have been encountered before, but no obvious solutions.
>>> >> Any
>>> >> > ideas?
>>> >> >
>>> >> > Thanks,
>>> >> > Brian
>>> >> >
>>> >> > --
>>> >> > Brian Radak
>>> >> > Postdoctoral Appointee
>>> >> > Leadership Computing Facility
>>> >> > Argonne National Laboratory
>>> >> > _______________________________________________
>>> >> > AMBER mailing list
>>> >> > AMBER.ambermd.org
>>> >> > http://lists.ambermd.org/mailman/listinfo/amber
>>> >>
>>> >>
>>> >>
>>> >> --
>>> >> -------------------------
>>> >> Daniel R. Roe
>>> >> Laboratory of Computational Biology
>>> >> National Institutes of Health, NHLBI
>>> >> 5635 Fishers Ln, Rm T900
>>> >> Rockville MD, 20852
>>> >> https://www.lobos.nih.gov/lcb
>>> >>
>>> >> _______________________________________________
>>> >> AMBER mailing list
>>> >> AMBER.ambermd.org
>>> >> http://lists.ambermd.org/mailman/listinfo/amber
>>> >>
>>> >
>>> >
>>> >
>>> > --
>>> > Brian Radak
>>> > Postdoctoral Appointee
>>> > Leadership Computing Facility
>>> > Argonne National Laboratory
>>> >
>>> > 9700 South Cass Avenue, Bldg. 240
>>> > Argonne, IL 60439-4854
>>> > (630) 252-8643
>>> > brian.radak.anl.gov
>>> > _______________________________________________
>>> > AMBER mailing list
>>> > AMBER.ambermd.org
>>> > http://lists.ambermd.org/mailman/listinfo/amber
>>>
>>>
>>>
>>> --
>>> -------------------------
>>> Daniel R. Roe
>>> Laboratory of Computational Biology
>>> National Institutes of Health, NHLBI
>>> 5635 Fishers Ln, Rm T900
>>> Rockville MD, 20852
>>> https://www.lobos.nih.gov/lcb
>>>
>>> _______________________________________________
>>> AMBER mailing list
>>> AMBER.ambermd.org
>>> http://lists.ambermd.org/mailman/listinfo/amber
>>>
>>
>>
>>
>> --
>> Brian Radak
>> Postdoctoral Appointee
>> Leadership Computing Facility
>> Argonne National Laboratory
>>
>> 9700 South Cass Avenue, Bldg. 240
>> Argonne, IL 60439-4854
>> (630) 252-8643
>> brian.radak.anl.gov
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
>
>
>
> --
> -------------------------
> Daniel R. Roe
> Laboratory of Computational Biology
> National Institutes of Health, NHLBI
> 5635 Fishers Ln, Rm T900
> Rockville MD, 20852
> https://www.lobos.nih.gov/lcb



-- 
-------------------------
Daniel R. Roe
Laboratory of Computational Biology
National Institutes of Health, NHLBI
5635 Fishers Ln, Rm T900
Rockville MD, 20852
https://www.lobos.nih.gov/lcb
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Nov 30 2016 - 11:30:04 PST
Custom Search