From: Francis E Reyes
Date: 22 October 2011 02:00
Hi all
I'm using dmmulti and one of my crystal forms seems to have a better solvent mask (after outputting it with solout) than the other. (Probably due to better phased reflections at low res for the first crystal form). Is it possible to take the better solvent mask and use it as input for the solvent mask for the other? How's this done? (a direct input of the better xtal1 mask to xtal 2 gives 'inconsistent cell info').
Thanks!
F
---------------------------------------------
Francis E. Reyes M.Sc.
215 UCB
University of Colorado at Boulder
----------
From: Pete Meyer
Hi,
It's definitely possible, but the devil is in the details. It's a two-step process: transform the coordinates of the mask so it's in the right spot in the new cell, and interpolate the mask into the appropriate grid for the new crystal form.
The last time I did this I was using O/PHASES masks (so I'm probably off on the details for CCP4 stuff): maprot should be able to handle the transformation and interpolation, but you may have to use a mask -> map -> maprot -> mask pathway. This isn't ideal due to differences in mask interpolation vs map interpolation but could be a good starting point.
Determining the transformation may be tricker - I'd try placing pseudo-atoms at similar features in both maps and using lsqkab to get the transformation matrix; but this approach would probably need tuning to get a good alignment.
Good luck,
Pete
----------
From: Francis E Reyes
Hi Pete
Thanks for the reply.
As for the transformation, I was going to use the refined mask operator (either from the end of a DM run that produces a better map for features I know have to be there, or from a correlation of the averaging mask from MAVE).
The interpolation is going to be a little more difficult. The 'good' solvent mask output after a dmmulti run
has different gridding, extent, and cell than the mask output by the other crystal forms.
However, the documentation for dmmulti says that SOLin can be on any grid or axis ordering.. not quite sure about the extent.
Furthermore, it seems that Poul Nissen was able to put in a generic mask for all his crystal forms (http://journals.iucr.org/d/issues/2010/03/00/kw5018/index.html) without regard for interpolation (unit cell, extent, or gridding) nor transformation. So I'm not exactly sure how I'm getting an 'inconsistent cell info'.....
Maybe it's because I'm getting my masks from dm instead of generating it from a my averaging mask (which came from a pdb).. hmmm.......
Anyone know of a wiki page that talks about maps/masks and a description of their properties (gridding, extent, basically everything discussed in the output of mapinfo?) A wiki with examples (images) of different settings for these items would be especially useful.
F
----------
From: Guenter Fritz
Hi,
have a look into the supplement of the paper by Poul Nissen pointed out.
There are some shell scripts exactly for that purpose.
I had to adapt the scripts a bit, but the instructions are very well to follow.
HTH
Guenter
Date: 22 October 2011 02:00
Hi all
I'm using dmmulti and one of my crystal forms seems to have a better solvent mask (after outputting it with solout) than the other. (Probably due to better phased reflections at low res for the first crystal form). Is it possible to take the better solvent mask and use it as input for the solvent mask for the other? How's this done? (a direct input of the better xtal1 mask to xtal 2 gives 'inconsistent cell info').
Thanks!
F
---------------------------------------------
Francis E. Reyes M.Sc.
215 UCB
University of Colorado at Boulder
----------
From: Pete Meyer
Hi,
It's definitely possible, but the devil is in the details. It's a two-step process: transform the coordinates of the mask so it's in the right spot in the new cell, and interpolate the mask into the appropriate grid for the new crystal form.
The last time I did this I was using O/PHASES masks (so I'm probably off on the details for CCP4 stuff): maprot should be able to handle the transformation and interpolation, but you may have to use a mask -> map -> maprot -> mask pathway. This isn't ideal due to differences in mask interpolation vs map interpolation but could be a good starting point.
Determining the transformation may be tricker - I'd try placing pseudo-atoms at similar features in both maps and using lsqkab to get the transformation matrix; but this approach would probably need tuning to get a good alignment.
Good luck,
Pete
----------
From: Francis E Reyes
Hi Pete
Thanks for the reply.
As for the transformation, I was going to use the refined mask operator (either from the end of a DM run that produces a better map for features I know have to be there, or from a correlation of the averaging mask from MAVE).
The interpolation is going to be a little more difficult. The 'good' solvent mask output after a dmmulti run
has different gridding, extent, and cell than the mask output by the other crystal forms.
However, the documentation for dmmulti says that SOLin can be on any grid or axis ordering.. not quite sure about the extent.
Furthermore, it seems that Poul Nissen was able to put in a generic mask for all his crystal forms (http://journals.iucr.org/d/issues/2010/03/00/kw5018/index.html) without regard for interpolation (unit cell, extent, or gridding) nor transformation. So I'm not exactly sure how I'm getting an 'inconsistent cell info'.....
Maybe it's because I'm getting my masks from dm instead of generating it from a my averaging mask (which came from a pdb).. hmmm.......
Anyone know of a wiki page that talks about maps/masks and a description of their properties (gridding, extent, basically everything discussed in the output of mapinfo?) A wiki with examples (images) of different settings for these items would be especially useful.
F
----------
From: Guenter Fritz
Hi,
have a look into the supplement of the paper by Poul Nissen pointed out.
There are some shell scripts exactly for that purpose.
I had to adapt the scripts a bit, but the instructions are very well to follow.
HTH
Guenter
No comments:
Post a Comment