Next steps: building into density

[Tristan Croll]

A lovely example I came across the other day:

open 3esu structureFactors true
view /F:1140; cofr center
delete /F:1145,1140,1246,1372,1371,1311,1310

What's there is clearly a HEPES molecule from the buffer, which was obscured by an overzealous water-picking algorithm adding waters to most of the peaks. In the near-term future, I'd really like to start providing tools for the the user to build into density like this (as well as waters, ions, and the peptide/nucleic acid chain itself, of course).

A few things that would make this much easier:

• Python access to residue templates (#1682)

• The map segmentation tool from Chimera. The more the target density can be isolated from its surroundings, the easier everything should become.

From there, the rest should be mostly straightforward. Fetch the template from the CCD, add the residue from the template, do some initial rigid-body fit and perhaps some degree of (non-MD) flexible docking, then settle into place using a simulation (initially in isolation from the surroundings if appropriate).

[goddard@…]

I have been porting the Segger segmentation tool from ChimeraX.  Good quality segmentation is rarely possible, so you should not expect it is going to simplify model building much.  Really it is an extra complication, for example when the segmentation bounds cut through the asymmetric unit.

[goddard@…]

Meant porting Segger from Chimera to ChimeraX.

[tic20@…]

In that case, the best approach in Cryo-EM density would probably be to generate a map with the model contribution subtracted. I have all the tools for that available to me, but for best results I’ll need to have B-factor refinement implemented first (something that’s been on my to-do list for a while).
 

 

[Eric Pettersen]

Seemingly all the requirements of this ticket are now satisfied.