Select 2D Thumbnail bug

998
request_recorded

#1

Hi after updating to the V2.13.2 version I’m noticing that the 9 thumbnails of the particles in ‘particles_selected’ and ‘particles_excluded’ on the right hand outputs are always the same. I can send a screen shot if it’s helpful

Bryan


#2

Hi @hansenbry,

Screenshots and the output of your “Overview” tab in the Select 2D job would be great, thank you!


#3

Here’s an example of the 2D select and I zoomed in on the Outputs field I was talking about. The particles_selected and the particles_excluded tiles are always they same since the update


#4

Hey @hansenbry,

Thanks for pointing this out, this has been fixed and will be released in the next update.

While we’re here, do you think it’d be useful to see more information about the classes you’ve selected?
For example: image of class, 1 image of a particle from the class, total number of particles in the class, class resolution
Also, do you think it’d be useful to see more images of the particles you’ve selected/excluded?


#5

Hi @sarulthasan

I think that seeing the number of particles and resolution in the 2D classification step is good enough for me. Having it again in the 2D selection step wouldn’t be needed, but the ability to see what the particles in a class look like would be very helpful. That way I could see if a class is bad due to alignment issues, ice thickness, or “bad” particles.

Bryan


#6

Great, how many particle images per class do you think will be necessary to answer these questions?


#7

I agree this would be helpful. Maybe a selection of 9 of the highest confidence particles from the class?

The other thing that would be very helpful would be if cryoSPARC ordered the 2D classes by similarity, and aligned similar classes so they are not in arbitrary orientations. This would make it much easier to compare different side view classes, for example.

Cheers
Oli


#8

It would be nice to see all of them in some type of pop out window or to export as an mrc stack and view in your viewer of choice. I think at least the top 50 particles would be good though.


#9

Hi @hansenbry, @olibclarke,

Thank you both for your input!