[Chimera-users] multiscale models and morph conformations
brignole at scripps.edu
Tue Jun 24 14:11:19 PDT 2008
So now I've got pdb's fit into my maps (using the excellent volume segmentation by color, volume eraser, and fitting tools!) and I'm trying to model the conformational changes. The fit pdb's for each EM structure (all saved into one large pdb with submodels that can conveniently be repositioned relative to one another) have exactly the same chains just with rearranged positions of the original domains, but submodel 0.1 of each model had two chains so morph conformations gets confused cause the rest of the submodels 0.2-12 only had one chain. Morph-conformations works fine if I save my pdb as mol2, then there are no longer submodels. That's not such a pain, except that mol2 doesn't preserve residue number so displaying particular residues becomes difficult.
So now I am trying to display EM-like structures in motion by using multiscale models of the intermediate conformations but the multiscale model doesn't follow its chain. So I suppose at this point the only way to visualize this is to write out each of the 60 intermediate pdb's, then through some scripting open each individually, make the multiscale model, display appropriately, save image. And then use ImageMagick to string these images into an animation to visualize the Morphing conformations. Also I can easily color the atoms of different chains but I'm having difficulty applying these colors to the multiscale models. Oops, wait I just discovered that MultiScale.color_surfaces_to_match_atoms() only works if I first select chains! Excellent. Still need some advice on multiscale movies though.
Many thanks for all your help.
PS. The problem with findlevels wasn't that the map was inverted...I arrive at negative thresholds because I hadn't set the voxel size correctly. Doh! All goes well after that.
The Morph Conformations tool works with submodels. It will not work
between two models or submodels that have different numbers of chains.
The error message you get means that you are giving it models with
different numbers of chains.
Edward Brignole wrote:
> ... I'd like to use morph conformations between two pdbs each
> composed of the same submodels/chains. I've tried to do this through the
> dialog but it displays an error telling me "cannot interpolate the
> models have different numbers of chains". The only way I can see to do
> this is to write the pdb as a mol2 so that all atoms are in a single
> model without submodels, open the mol2, and run morph conformations.
> Is there another method or is this the simplest?
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 4142 bytes
Desc: not available
More information about the Chimera-users