I'm a new user for Relion. Sorry if my question has been asked before. 

My data contains 80% of one molecules and about 20% ribosome. During picking using Eman boxer, it cannot be distinguish between this two molecules although it's rather easy by eye to pick out. However, I want to keep it as automatic picking since we want to start a huge dataset later on.

During the 2d classification, I eliminate the bad class already but probably ribosome particles are still present. 

The molecule I'm interested in has D3 symmetry.  In the 3d classification, I generate 5 classes with imposed D3 symmetry. One looks good, the other 4 look bad with artifact from symmetrization.

I wonder whether I should run 3D classification without D3 in this case and refine later with D3?

By the way, if I run the data initially with bin 2, how can I transfer the coordinate to start a local refinement with bin 1?

When will Relion-1.3 be released? The picking procedure described seems to be very fit for my purpose.

Best,

Huy