Blob picker missing highest contrast particles (negative stain)

[olibclarke]

When we process negative stain data, importing with the negative stain flag on, the blob picker seems to pick around the particles, missing most of them.

It seems like this has been reported before:

Blob Picker picking around particles in v2.13.2

I’m trying to do a negative stain reconstruction, with v2.13.2 but I’m having an odd issue where the blob picker and even the template picker seem to be picking around my white particles in the image. Is there a secondary setting or something other than selecting negative stain at input that I should be doing?

Negative stain - blob picks dark-on-light & 2D class light-on-dark

My negative stain images don’t pick well (with blob picker) when I toggle on negative stain when importing micrographs. But the blob picker does work well when I don’t toggle on negative stain. Once I’ve gotten locations with the blob picker, I’d like to do 2D class, but with negative stain toggled on. How would I do this? I tried importing the micrographs twice, with and without negative stain toggled, and then combining this with the particle locations. I tried exporting particles after pi…

Was there a resolution to this issue? What is the recommended workflow for processing negative stain data in cryosparc?

Cheers
Oli

[olibclarke]

After Inspect picks, it is clearly picking some particles, and it seems to me that the ones it is missing are those with the highest contrast. Perhaps there are some automatic exclusions based on the local power that are being applied behind the scenes, which are appropriate for cryo but not for negative stain? This is with the upper power threshold set to the maximum value:

[DanielAsarnow]

Oli, does it work if you just flip the micrographs first, e.g. e2proc2d.py --mult=-1 ... ?

I’ve always found cryoSPARC to work perfectly with negative stain data, if I just flip the micrographs first…

[olibclarke]

I have not tried that but will give it a go, thanks @DanielAsarnow!

[spunjani]

Hi @olibclarke, when you have a chance, please let us know if flipping the micrographs per @DanielAsarnow’s suggestion, fixed this for you

[olibclarke]

Hi @spunjani - I will certainly try this and I am sure it will work, but this is a workaround - if the “negative stain” mode at micrograph import doesn’t work, it should be fixed or removed, I think.

[apunjani]

Hi @olibclarke @DanielAsarnow we’ll triple check that we don’t have a missing negative sign

Btw on some negative stain data I found that setting the Lowpass filter in picking (blob or template) to low resolutions (eg. 120A) helped to some degree

[olibclarke]

Hi @apunjani, to me it seems to be an issue of some kind of contrast threshold, not a missing negative sign - because it is correctly picking some of the particles, just missing those with the highest contrast, as far as I can tell. Whereas if I import NS data as cryo, it picks around the particles and doesn’t pick any negatively stained particles at all.

Cheers
Oli