Hello Dr. Roe,
Thank you for your response, this has solved a lot of the confusion I was having about the results I was getting.
Frame 60521 with the pdb co-ordinates of 0.000 that I was referring to was taken from the original trajectory. As you said, it is likely corrupt. This is very likely since the mdcrd file was written in ASCII format. I should have written it as a netcdf trajectory (keyword ioutfm=1), and will do this in future runs. The Amber 12 manual also advises using iwrap=1 to keep the coordinate output from overflowing the trajectory files for long runs, and I will use this as well.
Thank you for sharing your cpptraj stripping and RMSD protocols. I tried this and it is much faster than what I was doing.
Cassandra
On 2013-02-06, at 11:34 AM, Daniel Roe wrote:
> Hi,
>
> On Wed, Feb 6, 2013 at 10:58 AM, Cassandra Churchill
> <churchil.ualberta.ca> wrote:
>> Following the simulation of a system for 15ns (trajectory file of 75 000 frames), I am attempting to analyze the trajectory with an RMSD calculation with ptraj. After reading in the trajectory files and striping the solvent and ions, I use the commands "center .CA mass origin" and "image origin center, and then calculate the mass weighted RMSD with respect to the first frame in 3 ways: (1) of C-alpha atoms, (2) of the protein backbone, (3) of all atoms.
>
> You have to be very careful when combining 'strip', 'image' and
> 'rmsd'. Without seeing your exact input to ptraj I can't make specific
> recommendations, but in general this is the procedure I follow: First
> strip, image, and output the stripped trajectory for analysis. The
> reason to perform imaging before or during a strip phase is because
> imaging requires box information, which you typically get rid of when
> stripping. Also, I recommend you use cpptraj and 'autoimage' for
> imaging, since imaging in ptraj can require extensive care to avoid
> imaging artifacts. For example:
>
> Step1 (cpptraj):
> parm prmtop
> trajin mdcrd.nc
> strip :WAT,Na+,Cl- outprefix nowat
> autoimage
> trajout nowat.mdcrd.nc netcdf nobox
>
> This will generate an imaged/stripped trajectory with no box info
> named 'nowat.mdcrd.nc' and a corresponding topology named
> 'nowat.prmtop'. This has the added benefit that subsequent analysis
> using 'nowat.mdcrd.nc' will be much faster. After I have a stripped
> and imaged trajectory, then I perform any analysis (like RMSD):
>
> Step2 (cpptraj):
> parm nowat.prmtop
> trajin nowat.mdcrd.nc
> rms Calpha .CA first out rms-CA.dat
>
>> To further attempt to isolate this problem, I obtained the pdb for frame 60521 from the trajectory, and some of the coordinates in this pdb are 0.000.
>
> If this is from the original trajectory, the frames in question are
> likely corrupted. If you're using netcdf trajectories you can safely
> discard corrupted frames and perform analysis on the non-corrupted
> frames. However, if you're using ascii trajectories it's much harder
> to recover frames after corruption occurs.
>
> -Dan
>
> --
> -------------------------
> Daniel R. Roe, PhD
> Department of Medicinal Chemistry
> University of Utah
> 30 South 2000 East, Room 201
> Salt Lake City, UT 84112-5820
> http://home.chpc.utah.edu/~cheatham/
> (801) 587-9652
> (801) 585-9119 (Fax)
>
> _______________________________________________
> 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 Thu Feb 07 2013 - 10:00:03 PST
