Hello all,
I am trying to process dataset of 90 kDa helicase. I have reported following parameters used for 2d classification. Particles were extracted using 256 box size. Pixel size of the micrograph was 0.74 A and particles were picked by circular blob and size was 70-160. I would like to know any changes in strategy while picking or processing so that classes will be clear.
Window dataset (real-space): true
Window inner radius: 0.85
Window outer radius: 0.99
Number of 2D classes: 80
Maximum resolution (A): 6
Maximum alignment res (A): null
Initial classification uncertainty factor: 1
Use circular mask on 2D classes: true
Circular mask diameter (A): 100
Circular mask diameter outer (A): null
Re-center 2D classes: true
Re-center mask threshold: 0.2
Re-center mask binary: false
Align filament classes vertically: false
Output filament in-plane rotations: false
Remove duplicate particles: true
Minimum separation distance (A): 20
Micrograph pixel size (A): null
Use clamp-solvent to solve 2D classes: false
Do CTF correction: true
Minimum alignment res (A): null
Sort classes by number of particles: true
Plotting sort method: similarity
Allow templates to shift in similarity sorting: false
Hard classify for last iteration: false
Do orientation alignment: true
Force max over poses/shifts: true
CTF flip phases only: false
Randomly perturb poses radians: null
Number of final full iterations: 1
Number of online-EM iterations: 40
Batchsize per class: 400
2D initial scale: 1
2D zeropad factor: 2
Ignore DC from image data: true
Min over scale after first iteration: false
Enforce non-negativity: false
Use FRC based regularizer: true
Use full FRC: true
Iteration to start annealing sigma: 2
Number of iteration to anneal sigma: 15
Use white noise model: false
Show plots from intermediate steps: true
Show scale bars: true
Random seed: 500322462
Cache particle images on SSD: false
Number of GPUs to parallelize: 1
Keep results of every full iteration: false
