[gmx-users] REMD demuxing

[Andrea C. Vaiana]

Hello everyone,
I'm trying to find a better way of demuxing REMD runs, I wrote a python 
script to do this but it is very slow because it relies on system calls 
to trjconv in order to read/write frames. I checked the documentation 
and mailing lists, it seems that this should be possible using "trjcat 
-demux...". I've tried this and reached similar problems (seemingly 
unresolved...) as Merc Martens in july. I've attached the last mail from 
that thread at the bottom of this email.

I have an remd.xvg file that looks like this:

0.000000   0   1   2   3   4   5   6   7   8   9   10   11   12   13   
15   14   16   17  ....
0.187500   0   1   2   3   4   5   6   7   8   9   10   11   12   13   
15   14   16   17
0.375000   0   1   2   3   5   4   6   7   8   9   10   11   12   13   
15   14   16   17
0.562500   0   1   2   3   5   4   6   7   8   9   10   11   12   13   
15   14   16   17
0.750000   0   1   2   3   5   4   6   7   8   9   10   11   12   13   
15   14   16   17
0.937500   0   1   2   3   5   4   6   7   8   9   10   11   12   13   
16   14   15   17
.
.
.



When I run:
trjcat -dt 0.75 -f  run3.0.xtc ... run3.47.xtc -o  run3.U.0.xtc ... 
run3.U.47.xtc   -demux remd.xvg

NOTE: I dump to the xtc file every 0.75ps but I'm using an xvg file 
(created from my log file) with all the replica exchanges i.e., every 
0.1875ps. Would this cause problems?

I get 48 identical xtc files named: run3.U.0.xtc #run3.U0.xtc.1# ... 
#run3.U0.xtc.47#
each of these is very small (219 K wheras my trajectories are 10 M).

a gmxdump on one of these produces:

run3.U.0.xtc frame 0:
   natoms=         0  step=         0  time=         0  prec=2.8026e-45
   box (3x3):
      box[    0]={ 5.65765e+00,  0.00000e+00,  0.00000e+00}
      box[    1]={ 0.00000e+00,  5.44156e+00,  0.00000e+00}
      box[    2]={ 0.00000e+00,  0.00000e+00,  5.42684e+00}
not available: x
run3.U.0.xtc frame 1:
   natoms=         0  step=       500  time=      0.75  prec=2.8026e-45
   box (3x3):
      box[    0]={ 5.65765e+00,  0.00000e+00,  0.00000e+00}
      box[    1]={ 0.00000e+00,  5.44156e+00,  0.00000e+00}
      box[    2]={ 0.00000e+00,  0.00000e+00,  5.42684e+00}
not available: x
run3.U.0.xtc frame 2:
   natoms=         0  step=      1000  time=       1.5  prec=2.8026e-45
   box (3x3):
      box[    0]={ 5.65765e+00,  0.00000e+00,  0.00000e+00}
      box[    1]={ 0.00000e+00,  5.44156e+00,  0.00000e+00}
      box[    2]={ 0.00000e+00,  0.00000e+00,  5.42684e+00}
not available: x
.
.
.

Does anyone know what's wrong?
Alternatively: does anyone have a fast python routine for 
reading/writing xtc frames?


thanks for any help.

Andrea

> merc mertens
> Mon, 17 Jul 2006 09:20:41 -0700
>
> hello,
>
> i am still trying to figure out how demultiplexing with trjcat (version 3.3.1) 
> works.
> i produced 3 trajectories, each containing only one frame:
> e.g.
>
> traj0.xtc frame 0:
>    natoms=     14670  step=       500  time=         1  prec=      1000
>    box (3x3):
>       box[    0]={ 8.47401e+00,  0.00000e+00,  0.00000e+00}
>       box[    1]={ 0.00000e+00,  4.98164e+00,  0.00000e+00}
>       box[    2]={ 0.00000e+00,  0.00000e+00,  3.48938e+00}
>    x (14670x3):
>       x[    0]={ 1.95100e+00,  2.11100e+00,  3.09600e+00}
>       x[    1]={ 1.95000e+00,  2.14800e+00,  3.19000e+00}
>       x[    2]={ 1.88400e+00,  2.15100e+00,  3.03200e+00}
>       x[    3]={ 1.94800e+00,  2.01100e+00,  3.11200e+00}
> ...
>
> traj1.xtc and traj2.xtc look the same (apart from containing different 
> coordinates).
>
> remd.xvg for -demux contains only one line:
>
> "1 2 1 0"
>
> using those files with:
> trjcat -f traj0.xtc traj1.xtc traj2.xtc -demux remd.xvg
>
> produces: trajout.xtc
>
> trajout.xtc frame 0:
>    natoms=         0  step=       500  time=         1  prec=  -1.99104
>    box (3x3):
>       box[    0]={ 8.48018e+00,  0.00000e+00,  0.00000e+00}
>       box[    1]={ 0.00000e+00,  4.98527e+00,  0.00000e+00}
>       box[    2]={ 0.00000e+00,  0.00000e+00,  3.49192e+00}
> not available: x
>
> so, even if there are definitely coordinates in the trajectory and there can 
> not be a mistake in the frame number, since there is only one, it seems strange 
> to me, that trjcat does not output anything but the box to the *xtc file. 
> further, the statement "not available: x" seems curious to me. 
>
> thanks for any help,
> merc
>
>
>
>
>
>
>
>   

-- 

Andrea C. Vaiana
Post Doctoral Research Associate
Theoretical Biology and Biophysics
Mail stop K710, T-10
Los Alamos, NM 87545
Los Alamos National Laboratory