[3dem] help with volume alignments

Fri Jul 25 07:58:31 PDT 2014

Dear 3DEM community,

Well you guys have made my weekend! Yesterday all was grey, doom and gloom, misaligned maps, today the sun is shining and thanks to you I have lovely aligned maps and nice looking FSC curves.

I'll just summarize the feedback for anyone needing to do this in the future.

What worked really quickly and easily was to use pdb2mrc (EMAN1) to convert the fitted pdb to an mrc file (thank you Justin Kollman), this kept the relative orientation in the map. Then I used the great new EMDB web server to do the FSC comparison (http://www.ebi.ac.uk/pdbe/emdb/validation/fsc/<https://urldefense.proofpoint.com/v1/url?u=http://www.ebi.ac.uk/pdbe/emdb/validation/fsc/&k=AjZjj3dyY74kKL92lieHqQ%3D%3D%0A&r=U7zQGZLkWJmK8RMBK%2B9tL7nEXHc%2FXErvXuSxcy9E1CI%3D%0A&m=8TFir7bCCEAgfCy%2BVPQP1SkxTnF%2Bdvku1PCiYj59QxQ%3D%0A&s=46e22165ad3fb9e08dfa5b767435cbc54a59d936b7690dcf7af9065f32e2ed58> - thanks Huy Bui for the link).

Carlos Oskar pointed me towards the conversion in xmipp (xmipp_volume_from_pdb, with a volume alignment if needed xmipp_volume_align, (--frm option for fast alignment))

Pawel Penczek pointed me towards these operations in SPARXS which preserve the orientations on conversion - SPARX wiki, and very helpfully (this is something I've been wanting to know for a while for other stuff) - any transformations done in Chimera and stored in its command line interface can be converted to Eulerian angles of SPARX or SPIDER, using http://sparx-em.org/sparxwiki/align_chimera

Ambroise Desfosses gave me feedback on how to resample a molmap stimulated map on the grid of the original map, which then preserves orientations on conversion to spider.
-open your experimental EM map (let's say it will be modelID #0)
-open your pdb and fit it.
-use molmap to make a map out of your pdb, using Nyquist as resolution filter. Say now your molmap map is modelId #2
-re-sample this simulated map :
vop resample #2 onGrid #0
-save this new mrc map.

Gregory Alushin told me how to do this in EMAN2: e2proc3d a.mrc --calcfsc b.mrc fsc_out.txt, pointing out that a problem might be some ringing at high spatial frequencies due to edge artifacts which can be overcome by:
A problem with this simple approach is that you may get "ringing" at high spatial frequencies due to edge artifacts. If this happens two potential ways of addressing it would be:

a) to soft mask both of your inputs, or

b) only calculating the fsc within a certain region of the two volumes, which also requires a binary mask

If your complex is globular a sphere would be a very reasonable mask choice.

Thanks again to everyone that helped me out,
All the best,
Kat

On Jul 24, 2014, at 7:19 PM, Toropova, Ekaterina wrote:

Dear 3DEM community,

I've been struggling all day with this problem and admit defeat, would love some help.

I'm trying to do an FSC comparison between my map and a crystal structure, my stumbling block has been to keep the two maps aligned. This is what I've tried.

First approach:
- I've docked the pdb file into my map in Chimera and saved it in the coordinate system of the map.
- I've then used the 'cp from pdb' command in spider to convert the pdb to a spider volume. And here's where I've got a problem - this introduces a rotation, such that when you open the volume in Chimera, it no longer aligns with the EM map.
- I thought I had got around this with a spider RT 3D rotation of 270,180,0 (phi,theta, psi). As it turns out this got me most of the way there - enough to fool me into proceeding with the FSC but on getting lower values than I was expecting I went back to look at the fit and saw that there was still a small rotation and possibly shift that I have spent the last few hours trying to figure out with no avail.

Second approach:
I then decided to try to do this in Chimera using the molmap command to convert the fitted pdb to density, save the map as an mrc file to later convert to spider. This does work - the map stays in the same place as an mrc but when I convert to spider the maps are no longer aligned, there's some discrepancy in the volume center that I can't figure out.

So my questions are:
1) Does anyone know the rotations that need to be applied to a volume generated from the 'cp from pdb' spider command to keep exactly the same orientation as the original pdb when viewed in Chimera?
2) Can I do the FSC comparison with mrc files?

Thanks very much in advance,

Kat

Katerina Toropova
Charles H. Hood Postdoctoral Fellow (Leschziner lab)
Dept. of Molecular and Cellular Biology
Harvard University
305 NWB, 52 Oxford St.
Cambridge, MA 02138

_______________________________________________
3dem mailing list
3dem at ncmir.ucsd.edu<mailto:3dem at ncmir.ucsd.edu>
https://mail.ncmir.ucsd.edu/mailman/listinfo/3dem

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.ncmir.ucsd.edu/mailman/private/3dem/attachments/20140725/ce2b58e3/attachment.html>