Hi Silvia,
You seem a little confused about what was written to your original mdcrd
file and what is in your new one. Let e try to explain what 'I think' you
have.
Firstly from your original MD simulation. This will have run for nstlim
steps, in your case what you suggest is 5 million which at 2 fs time step
would be 10ns. During this time the code will have written the coordinates
of ALL the atoms, including the water and ions, to the mdcrd file every ntwx
steps. Normally people set this in the range of 1000 to 2500 giving you
snapshots every 2 to 5ps. In your case though I suspect you set ntwx=50
which gave you a snapshot every 0.1ps and a final 60GB mdcrd file containing
100,000 frames. There is no way to recover the structure between these
frames since the software never wrote it to disk. It is unlikely that you
would ever require such fine grained storage for most applications.
At this point your trajectory file includes all the waters which makes it
very large. if you are not interested in the waters you can use ptraj to
strip them. This gives you the 5GB file, still containing 100,000 frames but
with just the protein and ions present. To load or visualize this trajectory
in VMD or in ptraj you will need a prmtop file that also has the water
stripped. RDPARM can produce such a file for you or you can generate one
using leap by saving the prmtop file prior to you solvating the system.
If you want to reduce the frame count further you can run this back though
VMD and give it a step size of something other than one.
E.g.
trajin foo.mdcrd 1 100000 10
trajout foo2.mdcrd
This will give you a file, foo2.mdcrd containing every 10th frame. You
snapshots are now 1ps apart and the file will contain a total of 10,000
frames and will be 500MB in size.
I hope this makes sense.
Alternatively you can use the option 'ntwprt' in the cntrl namelist to
specify the atoms, from 1 to X that will be written to the mdcrd file. E.g.
if you protein has 12,386 atoms specifying this number will result in the
mdcrd file containing only your protein. In order to post process this
trajectory file or load it into something like VMD you will require a
corresponding prmtop file that similarly has the atoms that were not written
stripped.
Good luck,
Ross
> -----Original Message-----
> From: amber-bounces.ambermd.org [mailto:amber-bounces.ambermd.org] On
> Behalf Of Silvia Carlotto
> Sent: Tuesday, November 03, 2009 8:43 AM
> To: AMBER Mailing List
> Subject: Re: [AMBER] 60 giga output file .mdcrd file
>
> Thanks to all,
>
> I reduce a 60 Giga file to a 5 Giga file with the commad strip in ptraj
>
> and the command generate 100000 set with no error.
>
> Is it possible to control in some way if all the 5 milions of step in
> the
> original
>
> file are all in the new small file?
>
> I don't understand if the number of set correspond to the nstep in some
> way.
>
>
>
> thanks
> _______________________________________________
> 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
Received on Tue Nov 03 2009 - 15:30:03 PST
