Blurry 3D classes: exclude or include?

3D Classification
1

Hi everyone,

After a first ab-initio job, I have a quite good subset of particles, which I would like to improve a little bit with a 3D classification step. I will try different resolutions to separate efficiently, but the result that I got made me asking a question:

J76_real_space_structures_iteration_60_online_em
J76_real_space_structures_iteration_60_online_em1170×975 147 KB

J76_real_space_difference_from_consensus_with_solvent_mask_iteration_70_online_em
J76_real_space_difference_from_consensus_with_solvent_mask_iteration_70_online_em1170×975 308 KB

Should I exclude the class 9 ? The shape seems ok, but I wonder why is it blurry and what is the meaning of this ?

Thank you for your advices.

2

It is misaligned particles. The shape is dictated by consensus refinement at beginning of job. Try NU-reform before the 3D class. Or exclude those particles though they may not be low quality. I also expect there are fewer particles in that class than the rest.

I would also suggest resolution 14 or so to sort heterogeneity at this stage

2 Likes