So, I *think* something like this is possible, but as Tom mentioned
there will have to be some bookkeeping on your part. Say you have a
100 frame long trajectory of 3 molecules, each 10 residues long,
taking up the first 30 residues of your system. First, you would use
'strip' to generate the individual copies of your molecules:
parm myparm.parm7
trajin mytraj.nc
strip !(:1-10) outprefix strip
outtraj Mol1.nc
run
strip !(:11-20)
outtraj Mol2.nc
run
strip !(:21-30)
outtraj Mol3.nc
run
Now you can read in the individual trajectories and perform your
clustering, using the 'summarysplit' and 'splitframe' keywords (only 1
parm since all molecules are identical):
parm strip.myparm.parm7
trajin Mol1.nc # Frames 1-100
trajin Mol2.nc # Frames 101-200
trajin Mol3.nc # Frames 201-300
cluster :
1-10.CA out cnumvtime.dat summary summary.dat info info.dat \
summarysplit split.dat splitframe 100,200 [<any other cluster options>]
Now you will have clustering results for the overall ensemble in
summary.dat and info.dat, results for each individual molecule in
split.dat, and the cluster number vs time (cnumvtime.dat) that you can
use to figure out which molecule is in which cluster in the original
trajectory for any subsequent analysis.
Hope this helps,
-Dan
On Fri, Jan 30, 2015 at 8:54 AM, 浅井 賢 <suguruasai.gmail.com> wrote:
> Dear tec3
>
> > Note that with either the closest or this option, you will still
> potentially have artifacts from the "holes" left over from the missing
> peptides which will be problematic if there is close peptide-peptide
> association...
>
> Yes. It is much better if I could cluster molecules *without* removing
> them and have complete trajectories of each clusters (I don't care if it
> is duplicated or not).
> That's why I asked this but it sounds not possible...
>
> > Note also, that investigating "dynamics" of water around a cluster is
> not defined since there would no longer necessarilly be equal time
> between frames
>
> I said 'dynamics' but actually it is not so sequential data is not
> required. But anyway thanks for your note :-)
>
>
> Regards
>
> Asai
>
> On 2015年01月31日 00:36, Thomas Cheatham wrote:
>>> Aha! I think I got it.
>>> 'closest' command modify coordinate frame *and topology* (not like
>>> 'strip') so write out the modified topology with 'parmwrite' and use the
>>> modified topologies to load the produced trajectories right?
>> ...while this will generate smaller trajectories, it will be rather
>> computationally demanding to generate... A simpler approach if you have
>> ample disk space is to simply strip out all the peptides except one, for
>> each peptide, keeping the water and ions. Note that with either the
>> closest or this option, you will still potentially have artifacts from the
>> "holes" left over from the missing peptides which will be problematic if
>> there is close peptide-peptide association... [Note also, that
>> investigating "dynamics" of water around a cluster is not defined since
>> there would no longer necessarilly be equal time between frames.]
>>
>> --tec3
>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
--
-------------------------
Daniel R. Roe, PhD
Department of Medicinal Chemistry
University of Utah
30 South 2000 East, Room 307
Salt Lake City, UT 84112-5820
http://home.chpc.utah.edu/~cheatham/
(801) 587-9652
(801) 585-6208 (Fax)
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Jan 30 2015 - 14:00:02 PST
