I processed a dataset of a sample less than 150 kDa recently. After template picking, I used different “Batchsize per class” with 200 and 500, respectively, to do 2D classification.
“Number of 2D classes” - 200 (also tried 100); “Initial classification uncertainty factor” - 4; “Force Max over poses/shifts” - OFF; “Number of online-EM iterations” - 100.
Most of 2D classes did not have clear feature. I chose good classes in each run. I thought there would be a large part of particles were duplicate. However, when I used “Remove Duplicate Particles”, a few particles (213; total particle number is ~580000) are removed. I am confused why this will happen. Does this mean the particles were not well aligned during 2D classification? Which parameters may I change to get better 2D results?
I attached chosen 2D classes for two runs here.