Contrast issues with drifty tilt series

[cla...@gmail.com]

Hi all,

I've recently been processing a cryo-FIB tilt series collection through Warp and I've discovered some interesting contrast issues. The motion correction of Warp has added a bright line at the bottom of many of the tilts that affects the contrast of the tilt series stack it generates (see below for an example). As such, the contrast of the area of interest is extremely poor and Aretomo then struggles to align the tilt series as it seems to focus on the bright area. Is there a way for Warp to generate the tilt series with a slight crop or for it to rescale the contrast, clipping the high values? Failing that, is there a way of using an external program to do this? I tried the -meansd in newstack and ccderaser but no joy so far. I've considered cropping the .st output but haven't in case the alignment .xf file is no longer valid for Warp, but if someone can confirm otherwise I can use that as a solution.

Cheers,

Mat

[Lorenzo Gaifas]

Hi Mat,

I've had similar issues with the contrast and AreTomo. Warp lets you choose the "unbinned tomogram dimensions" in the top left corner; you can simply change this to a slightly smaller size than the micrograph dimensions, and you;re good to go :) Warp reconstructs a box centered in the middle of the tomogram, so it will remove the same amount of pixels on all sides.

Best,

Lorenzo

--
You received this message because you are subscribed to the Google Groups "Warp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to warp-em+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/warp-em/1acb024d-1be1-4341-ba91-fb89759467adn%40googlegroups.com.

[cla...@gmail.com]

Hi Lorenzo,

Thanks for the suggestion, but I'm hoping to crop the tilt series (or minimise the contrast issue) rather than crop the final tomogram output. The edge effect is very dominant in most of my tomograms and cropping the final result doesn't help. Aretomo is working with some tilt series (showing the lines across the area) but struggling with others (poorly aligned with only lines visible). If there's no way around it, I do have Aretomo alignments generated using tilts corrected by MotionCor that don't have this issue but I have excluded more bad frames in Warp so I need to manually adjust my .xf files to compensate, which I'd rather not do for 100 tomograms if at all possible!

Cheers,

Mat

[teg...@gmail.com]

Hi Mat,

Is this from a K3? It might be fixable by using a defect map during tilt movie pre-processing as described here (relevant part copied below): http://www.warpem.com/warp/?page_id=185

Cheers,

Dimitry

Some camera sensors have small defective regions that can’t be corrected using a gain reference. They won’t affect alignments if they’re small, but can be annoying, especially in tomography. Warp can use a defect map to fill those regions with the average value of their surroundings. The defect map must have the same dimensions as the gain reference, and contain pixel values = 1 in the defect regions and = 0 everywhere else. The same flip X/Y and transpose operations will apply to the gain reference and defect map. Alternatively, you can have Warp create the defect map image based on a text description of rectangular regions. To do this, click Select defect map, change the file filter to *.txt, and select the text file. In this file, every line should describe a rectangle in the following format:

[bottom left corner X] [bottom left corner Y] [width] [height]

The X and Y coordinates start with 0, not 1. The columns are separated by spaces or tabs. For instance,

0 3709 3838 1

will mask the top-most row of a 3838×3710 px sensor. Once the file is loaded, Warp will generate the defect map image and save it under [project directory]/defectmap/defects.tif and automatically change the file path. Please check this image to make sure the mask is correct. For future projects using data from the same sensor, you can just reuse this TIFF file.

[cla...@gmail.com]

Hi Dimitry,

I gave this a go and it seems to be working, I still have some lines appearing in my final tomograms but the contrast is much clearer now and the tilt series alignment is now working. Thanks very much for your help!

Cheers,

Mat