RELION focused classification after M refinement

[毛翘顺（emmao）]

Hi everyone,

I'm relatively new to the Warp-RELION-M workflow, and I have a few questions about focused 3D classification after M refinement.

I currently have an M-refined reconstruction at ~4.64 Å resolution (Nyquist limit 3.88 Å). While reading some papers, I noticed that many workflows perform an additional refinement in RELION after M refinement before proceeding to focused classification.

This confused me a bit, because I thought the particles exported from M could be directly used for RELION 3D classification. Is there an advantage to doing another RELION refinement first?

For subsequent focused classification in RELION, would you generally recommend:

• no image alignment, or
• local angular searches (such as 0.9-1.8 degree)?

Also, after obtaining distinct classes from focused classification, is it recommended to take each class back into M for another round of refinement?

Finally, if a target state is relatively rare (ribosome in my case) and contains only a small number of particles (for example ~1,000–3,000 particles), would M still be suitable for refinement?

Thanks a lot for any advice, and apologies if these are basic questions — I'm still quite new to using Warp/M.

Thank you for any help!!

Best,
Qiaoshun