mask not being applied?

sstagg

mask not being applied?

Hi cisTEM folks,

I'm having trouble with the mask options with Manual Refine. If I understand correctly, when one chooses a user-defined mask, the mask is applied to the volume for a given iteration and placed in the Scratch/ManualRefine3D directory. However, when I look at the volume that directory, it doesn't look like the mask that I defined is being applied. Is this a bug, or am I not understanding how to use the options properly? If it's not a bug, is there a way to get it to apply a mask each iteration?

Thanks,

Scott

Tue, 07/09/2019 - 19:50

timgrant

Hi Scott,

As you say, the mask is applied just for the refinement, the input is the volume in the scratch directory that you specified, and so all the refinements are being done with a masked 3d. The reconstruction is not masked, as we think it's always best to have the unmasked reconstruction, and you can always mask it yourself (e.g. using the apply_mask command line program in cisTEM) if you need a masked copy.

Tim

Tue, 07/09/2019 - 19:53

sstagg

In reply to by timgrant

Hi Tim,

I think I'm not being clear. I agree that the output volumes shouldn't be masked, but I'm talking about the volume that is in the scratch directory, which I believe is what is used as input for the next refinement iteration. That would should be masked, right? When I look at it, I can tell that the volume has been masked in some way, just not the mask I am providing.

S

Tue, 07/09/2019 - 21:10

timgrant

Hi Scott,

Yes that should be masked with the mask you are providing.   That mask is interpreted as anything above 0 is 1, and anything 0 and below is 0.  It is then smoothed etc with the parameters provided in the expert options panel.

What sort of mask are you seeing?

Tim

Tue, 07/09/2019 - 21:20

sstagg

Mine looks like an automask of some sort, but my mask is not fully binarized. I thought cisTEM was just multiplying. I will try making the mask with just 0s and 1s and see if it helps.