Failed in read in DCD by the latest MSMBuilder

[Xiaojun Xu]

Hi All,

On webpage http://wiki.simtk.org/msmbuilder/User_Contributions, we found that "The DCD reader is already included in the latest MSMBuilder releases". However, even when I tried " ConvertDataToHDF.py -s ./dipeptie.pdb -I ~/msmbuilder2.0/Tutorial/DCD" to read the DCD format instead the XTC format from the XTC directory, it failed to find any data. Does this mean MSMBuilder2.0 still do not read any other trajectories than XTC? And if so, why does it include the DCD directory in there? Any answer or suggestions will be greatly appreciated!


Max Xu

[Alex Dickson]

Hi Max,

Try adding " --source=file_dcd " to your ConvertDataToHDF command.

-Alex

[Joshua Adelman]

Hi Max,

Looking at the source code, ConvertDataToHDF.py has the ability to read .dcd files. I believe your problem however is that you are not supplying the 'source' argument to the script, and the default will only look for .xtc files. So I believe (without testing things however), that if you add:

Code: Select all

--source file_dcd

or

Code: Select all

-S file_dcd

that it will work.

Also as a side note, I don't believe this is documented in the docs. Fortunately, MSMB2 is written in python which for the most part is fairly easy to read and understand. If you look in msmbuilder/Scripts/ArgLib.py you will see a complete list of all available options for the command line tools, along with a description and the default values. You can the entries there to the arglist that is in the __main__ section of a particular script:

Code: Select all

arglist=["projectfn", "PDBfn", "input", "source", "discard", "mingen", "stride", "rmsdcutoff" ]

Josh

[Xiaojun Xu]

Hi Alex, Hi Josh,

Thank you so much for your reply! I spent a whole day before I posted it. Seems like I should have done it earlier It gives me a heads up that I probably need to look into the scripts more carefully. Thanks again!

Cheers,

Max

[Joshua Adelman]

Before you follow my original advice to start digging through the source, I forgot that MSMB's command line tools are built around python's optparse module (I looked in the source since, when I answered you question, I was on a machine that I didn't have MSMB2 installed), so if you were to issue:

Code: Select all

ConvertDataToHDF.py -h

You would get a lit of all of the command line options and description, which includes:

-S SOURCE, --source=SOURCE
Data source: "file", "file_dcd" or "FAH". This is the
style of trajectory data that gets fed into
MSMBuilder. If a file, then it requires each
trajectory be housed in a separate directory like
(PROJECT/TRAJ*/frame*.xtc). If FAH, then standard FAH-
style directory architecture is required. Default:
"file"

[Xiaojun Xu]

This works great, saves me a whole lot of time! Thanks again, Josh!