Releasing tools for converting Leica Matrix Screener and LaVision Ultramicroscope files to Big Stitcher

Hi,

just a quick announcement that I’ve released two tools on Github that address data processing issues that arouse in our microscope facility, namely

  1. Converting tiled Leica Matrix Screener images to Big Stitcher/Big Data viewer files;
  2. Converting LaVision Biotech Ultramicroscope images to Big Stitcher/Big Data viewer files.

I created these tools because I was not able to import these files directly into Big Stitcher. Much of the heavy lifting under the GUI is done by the npy2bdv code by @nvladimus which in turn is based on imarispy by @talley.

There is lots of copy&pasted code between those two tools (might have made sense to factor out to common functionality into a separate package) and there is no guarantuee that it will work on your data set, but maybe it is useful for someone. Feel free to fork and modify as needed.

8 Likes

Awesome Volker! We will certainly use this in our CLEM pipelines.

@schorb , let me know if you run into any issues. In particular, I may have to add multi-channel support, for matrix screener but didn’t have a data set to test it (I try to stay away from the microscope these days).

Also, a use case not supported with the GUI (but the underlying python class) is to not just convert a single well but all wells into a single Big Stitcher project. I did not include this in the GUI as having that many tiles can cause issues. However, if you have the right display manager and get it to work it is rather impressive to see a whole well plate laid out in 3D with the ability to zoom in. It becomes very impractical to stitch, fuse and export though.

Great tools! Glad you found npy2bdv useful. Let me know if you need any new features implemented.

1 Like

@VolkerH, thanks a lot, a tool for Ultramicroscope will be useful for us.

@constantinpape maybe have a look whether your python code to produce bdv files could be merged?

Hi @VolkerH, I assume the main thing is to find the correct stage positions from the images and write them into the <affine> ... </affine> tag of the bdv.xml files? If that’s the case maybe it could be something that could be added to Bio-Formats, which has metadata files for stage positions (see here: Bio-Formats Image Position). Because if Bio-Formats presents there the correct values, probably BigStichter could directly read the files and you would not have to stage them.

@Christian_Tischer, finding the stage coordinates for the affine xml tags is indeed the main part. However, there are a few other things that would probably not work with the Big Stitcher import dialog in its current form, e.g. the nested folder structure of the Matrix screener files etc.

I agree, that this is something that could probably be addressed by changes to Bio-Formats and with some changes to Big Stitcher import dialogs. Somone has to do it though, I didn’t want to burden anyone else but required a solution quickly, so I chose tools I know how to work with.

1 Like

What do you think, would it make sense to get in touch with Frank Sieckmann and get a proper export option from MS straight away? This should not be so much of a struggle given that for their internal visualisation all data (incl. pyramids) is already there.

Martin

I totally understand and it is more than fair enough! I was just raising this point because I know now at least four people working in facilities having these kind of issues and I know that, in principle, the guys from Bio-Formats would be interested in helping. But, as you said it is a matter of resources. However, if enough people say they need this maybe one can also try to find resources… I will start another thread about this topic to see whether we can do something.

@schorb I actually think the output files from Leica Matrix screener are pretty easy to work with. The metadata is in there (probably more or less OME-tif conforming). The only thing that I could not find in there is the Z-drive positions and the Z-spacing (which could be derived from adjacent Z-drive positions anyway). See https://github.com/VolkerH/LeicaMatrixScreener2BigStitcher/blob/62d7014da35ba294ea2105cbf045ff2a779c4673/lm2bs/process_matrix_screener_data.py#L42
Sure it would be useful to have at least Z-drive positions in there (maybe I just overlooked them). For me, I don’t think I will be overrun with users using matrix screener at our facility, but if you have a use case it may be worth lobbying Frank in Monnem to put it in there. Somehow I doubt this will happen for older microscopes such as the SP5 though.

I had more issues with finding Metadata for the Ultramicroscope tif files and the weird filenaming.

@Christian_Tischer, if it worked just out of the box using Bioformats that would certainly simplify things.