Dear Julian,
The approach of using images recorded in a tilt series for 3D reconstruction using a single-particle cryo-EM approach is described in this paper.
Bartesaghi A, Lecumberry F, Sapiro G, Subramaniam S:
Protein secondary structure determination by constrained single-particle cryo-electron tomography.
Structure 2012, 20(12):2003-2013.
Sriram
On 4/26/17, 10:13 AM, "Julian Reitz" <[log in to unmask]> wrote:
Dear all,
I would like to give you a short description of what I want to do, so that you can understand my problems and hopefully help me to find a solution.
We are currently working on some cryoET data. The standard pipeline therefore would be to reconstruct the tomogram, find particles in the tomogram and make sub-tomogram averaging.
However our current tilt-series are difficult to align and due to the poor alignment we have a non optimal reconstruction and therefore also not the best resolution in the sub-tomogram averaging.
Our idea now was as follows:
First use the normal workflow. We then used the 3D coordinates of the particles in the tomogram to calculate back, where the projections of these particles are on the images of the tilt-series. We now wanted to use these information to proceed in a single-particle-like way in order to circumvent the problems with the alignment. Of course the particles we extract from the images in this way are not independent from each other, since we know the difference of the tilt angle from image to image. Is there a way to make use of these information with relion? I just tried to make a 3D-refinement (we have a very good starting model) but the result doesn't look good and I can see from the output star file that the orientations are not correct.
Any ideas?
Best,
Julian
|