Vesicle organelle masking

Sometimes it is beneficial to restrain the results from cross-correlation (template matching) either due to high correlation with artifacts or to limit results to a specific ROI within the tomogram. For example, one might search only for ribosomes within a mitochondria or wants to filter membrane-associated particles. pytom provides an option for applying a binary tomogram mask during candidate extraction or in the particle picking widget.

Requirements for the 3D mask:

• Tomogram/Score file dimensions need to be matched
• Binarization with ROI of the masked region set to 1 and background set to 0 (pytom will also try to binarize the mask, but this might give unexpected results)
• Supported formats are mrc (rec) and em.

Applying the mask:

• Option 1: in the pytomGUI Particle Picker open a tomogram and the particle list after template matching. Next browse for a mask via the bottom left and apply the mask to the currently loaded particle list. Save the masked particle list as a new file.
• Option 2: In extractCandidates.py provide the mask file via the --tomogram-mask flag

Tip: Custom binary 3D masks can be generated in avizo’s segmentation tool