Masking a local region of the map

[ashuthe...@gmail.com]

Dear all,

I am refining a single particle cryo em map, which currently goes to ~8A. There is an interface where I am particularly interested to know details of. I was thinking to use a mask covering that and near by region, excluding rest part of the map, for alignment to locally sharpen that area. This is only to get an idea of interacting residues at that interface by attempting to sharp the map locally, if possible. Is there a way to do this? With volume eraser tool of chimera, I could save the interested area alone, though I am not sure how to use this during refinement.

Thanks !

Ashu

[Steve Ludtke]

In the new version of EMAN2 (current snapshot or 2.2 test release) there is a --tophat option available in e2refine_easy. If you use --tophat=local, it will determine the local resolution of the map and filter accordingly, which will produce sharper features in the more self-consistent regions of the map. This would be the first thing I'd try. 

If the region you are interested in is a more flexible domain then either you need to classify based on that region, specified as a mask (e2refinemultinoali.py). Or you can try subclassification with alignment (e2refinemulti.py) to reduce the variability.  General discussion of these methods is here:

Ludtke, S. J. "Single-Particle Refinement and Variability Analysis in EMAN2.1." in Methods Enzymol 579159-189 (Elsevier, United States, 2016). PMC5101015

One other issue to ask yourself is why the resolution is limited? If you have nice thin ice, and collected the data on a Krios+K2, then I would say it's probably intrinsic flexibility of the specimen, which may be addressed computationally. If the data was collected on a 200 KeV LaB6 scope on a CCD camera, then the resolution may be intrinsically limited.  Just as an example...

 

----------------------------------------------------------------------------
Steven Ludtke, Ph.D.
Professor, Dept. of Biochemistry and Mol. Biol.                Those who do

Co-Director National Center For Macromolecular Imaging            ARE

Baylor College of Medicine                                     The converse
slu...@bcm.edu  -or-  ste...@alumni.caltech.edu               also applies
http://ncmi.bcm.edu/~stevel

--
--
----------------------------------------------------------------------------------------------
You received this message because you are subscribed to the Google
Groups "EMAN2" group.
To post to this group, send email to em...@googlegroups.com
To unsubscribe from this group, send email to eman2+un...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/eman2

---
You received this message because you are subscribed to the Google Groups "EMAN2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eman2+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[ashuthe...@gmail.com]

Thanks Steve,

I'll try tophat local, and then splitting the map and multi refinement if tophat doesn't work. I went through the section in your article which describes the theory of splitting the map. 

I have been discussing in the mailing list recently that I am kind of stuck with removal of bad particles. Data have been collected on Talos Arctica 200keV microscope, K2, direct detector, FEG etc. Ice wasn't thick but I had issue with particle picking, as not only there are ice contaminations, my protein is a monomer with predicted flexible linkers between globular domains. The protein size is ~300kD. Eventually, I manually picked 100K particles, made 2-D classes in viper and initial model in isac. Although there was a slope, there was no distinct distribution of good-bad particles  when I evaluated particles quality after refinement. Manual attempt to remove bad particles, removed ~50% of particles, which you suspect could be because of the lack of clear cut clustering. I am refining the map with 50% of 'good' particles. Although refinement is not complete, reported resolution seems to have increased from 7.7 (with 100k particles) to 7.2 A (with 50k particles). 

Most important thing to note here is that I don't see the density for approximately a domain of ~60kD ! This unseen domain is predicted to be attached through a linker of ~70 amino acids. I was thinking that probably this domain is highly flexible and given its size, it is unable to give a contrast enough to be seen in the images. Adding flexibility to the smaller size of this domain is probably the reason for its absence in the refined map. Although it is possible that rest of the 230kD region is very rigid (in terms of particle alignments). 

If okay with you,  I would be happy to share specific details with you personally.

Thanks so much

Ashu

[Steve Ludtke]

Hi Ashu. The bad particle exclusion method I discussed in EMAN2.2 based on the resolution range FRC plots really only works well when the particles are fairly homogeneous. ie - it is not useful in removing particles in the presence of conformational or compositional variability. If you have variability, you should deal with that first, by subclassification or other methods, then use the bad particle identification tool on a refinement done with the more homogeneous subset.   Anyway, yes, you are always free to communicate with me personally if there are details you are uncomfortable broadcasting.

----------------------------------------------------------------------------
Steven Ludtke, Ph.D.
Professor, Dept. of Biochemistry and Mol. Biol.                Those who do

Co-Director National Center For Macromolecular Imaging            ARE

Baylor College of Medicine                                     The converse
slu...@bcm.edu  -or-  ste...@alumni.caltech.edu               also applies
http://ncmi.bcm.edu/~stevel