[gmx-users] Re: AGAIN: demixing REMD trajectories

[madeleine.kittner at mpikg.mpg.de]

>>> madeleine.kittner at mpikg.mpg.de wrote:
>>>> Hi gromacs users,
>>>>
>>>> I worked a on my problem (see mail below) and got some new questions:
>>>>
>>>> I did REMD using 50 replicas using gromacs version 3.3, while the data
>>>> were saved every 20 ps and exchanges were attempted every 5 ps. I used
>>>> demux.pl to generate replica_temp.xvg and replica_index.xvg. To get
>>>> continuous trajectories I used trjcat from version 3.3.1 (it's not
>>>> available in older versions)
>>>>
>>>> trjcat -f traj*.xtc -demux replica_index.xvg
>>>>
>>>> I solved the segmentation fault problem by compiling gromacs 3.3.1
>>>> differently. But now I'm stuck with the same problem as other people
>>>> before. trjcat produces only one output file trajout.xtc which
>>>> contains
>>>> only the box parameters and looks like this:
>>>>
>>>>  trajout.xtc frame 0:
>>>>    natoms=         0  step=         0  time=         0  prec=
>>>> 0
>>>>    box (3x3):
>>>>       box[    0]={ 5.46920e+00,  0.00000e+00,  0.00000e+00}
>>>>       box[    1]={ 1.82296e+00,  5.15645e+00,  0.00000e+00}
>>>>       box[    2]={-1.82292e+00,  2.57794e+00,  4.46567e+00}
>>>> not available: x
>>>> trajout.xtc frame 1:
>>>>    natoms=         0  step=      5000  time=        20  prec=
>>>> 0
>>>>    box (3x3):
>>>>       box[    0]={ 5.46920e+00,  0.00000e+00,  0.00000e+00}
>>>>       box[    1]={ 1.82296e+00,  5.15645e+00,  0.00000e+00}
>>>>       box[    2]={-1.82292e+00,  2.57794e+00,  4.46567e+00}
>>>> not available: x
>>>> trajout.xtc frame 2:
>>>> .... and so on.
>>>>
>>>> Is the problem caused by the different time frames in the traj*.xtc
>>>> files
>>>> and the replica_index.xvg file? Or is it caused by mixing up the
>>>> different
>>>> gromacs versions? For other people using the demux.pl script seemed to
>>>> solve all the problems but it does not in my case.
>>>>
>>>> Thanks for your help.
>>>> Cheers, Madeleine
>>>>
>>> did you read the help text in the perl script?
>>
>> Yes, I did.
>>
>>> # If your exchange was every N ps and you saved every M ps you can make
>>> for
>>> # the missing frames by setting extra to (N/M - 1). If N/M is not
>>> integer,
>>> # you're out of luck and you will not be able to demux your
>>> trajectories
>>> at all.
>>
>> In my case N=5 ps and M=20 ps which means (N/M-1) is not an integer and
>> I
>> m unlucky. But I only need to know every 20 ps at which temperature each
>> replica is simulated. Even if I produce a replica_index file that
>> contains
>> only the information each 20 ps, trjcat produces one trajout.xtc file
>> which contains only the box parameter as posted above.
>
> I would think it is reverse N=20 M = 5, so you have to set the number to
> 3. You can test it by computing RMSD for the resulting trajectories and
> seeing whether these are continuous.
>
But this makes no sense to me. Then I get a replica_index file like this

0           0    1    2    3    4    5    ... x
1           0    1    2    3    4    5    ...
2           0    1    2    3    4    5    ...
3           0    1    2    3    4    5    ...
5           0    1    3    2    4    5    ... x
6           0    1    3    2    4    5    ...
7           0    1    3    2    4    5    ...
8           0    1    3    2    4    5    ...
10          0    3    1    2    4    5    ... x

The exchanges happen only every 5ps (x). I think this is a useful option
if you save your data more often than you exchange: like if I would save
data each 5 ps and exchange each 10 ps, than demux.pl would copy the line
0ps to line 5ps and so on.
I changed the output subroutine in the demux.pl script to write out
information about the replicas only each 20 ps, because that's the time
step in the traj files. I thought that would make a difference but it
doesn't. The output of trjcat is still only one trajout.xtc file like this

trajout.xtc frame 0:
   natoms=         0  step=         0  time=         0  prec=         0
   box (3x3):
      box[    0]={ 5.46920e+00,  0.00000e+00,  0.00000e+00}
      box[    1]={ 1.82296e+00,  5.15645e+00,  0.00000e+00}
      box[    2]={-1.82292e+00,  2.57794e+00,  4.46567e+00}
not available: x
trajout.xtc frame 1:
   natoms=         0  step=      5000  time=        20  prec=         0
   box (3x3):
      box[    0]={ 5.46920e+00,  0.00000e+00,  0.00000e+00}
      box[    1]={ 1.82296e+00,  5.15645e+00,  0.00000e+00}
      box[    2]={-1.82292e+00,  2.57794e+00,  4.46567e+00}
not available: x

There are no conformations saved. And shouldn't I get 50 trajout.xtc files?
What input exactly does trjcat require to use the -demux flag properly? I
thought the number of time frames in the traj.xtc files should match the
number of time frames in the replica_index file. Should the columns in the
replica_index file are spaced by tabs??

Hey thanks for your help so far, David!
Madeleine


>>> --
>>> David van der Spoel, Ph.D.
>>> Molec. Biophys. group, Dept. of Cell & Molec. Biol., Uppsala
>>> University.
>>> Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205. Fax: +4618511755.
>>> spoel at xray.bmc.uu.se	spoel at gromacs.org   http://folding.bmc.uu.se
>>

---
Madeleine Kittner, PhD student
Department of Theory and Bio-Systems
Max Planck Institute of Colloids and Interfaces
Research Campus Golm
14424 Potsdam, Germany