I am a beginner in Cryo EM. I have collected images of my protein which is about 150kDa and has a diameter of about 180A. While doing 2D classificiation, I am getting very poor classes. I have extracted the particles using box size 384→96. I am also attaching the 2D classification parameters for your reference.
Do you recognize particles on your micrographs, looking like what you would expect? Are you sure you have a good picking without too many false positives, and how many particles are going into your 2D classification?
If you share an example micrograph it might be easier to help, but from the 2D alone it looks like might have a lot of junk particles in your stack.
Seconding boggild, OP you might consider using exposure curation to test different parameters for your picking → 2D classification pipeline on 500-1000 of your micrographs. These averages seem to have a lot of junk particles still.
I often find it helpful to be more stringent in the beginning, until you have nice 2Ds to use for template picking or to curate small set of particles that can be used for ab initio models. From there you can pass a relatively large and “impure” particle stack over the ab initio models with heterogenous refinement.
It’s hard to say without seeing a micrograph, but you might get some benefit from setting “Maximum number of local maxima to consider” at template picking to a lower value (default is 4000, I often use 2000) if you are suffering from a large number of empty picks.
Thank you @EElliott and @boggild for your suggestions. I am expecting a V shaped structures and 278361 number of particles are going into the 2D classification. I have attached a reference micrograph for your reference.
