Preprocessing of imaging data

If you used our same experimental setup for imaging, you will obtain individual images as a result of performing imaging of the stained tissue in a tile-scan fashion. From this point, the open=ST pipeline expects to have a single image for the whole tile-scan, rather than individual files per tile.

This step depends on the microscope used for imaging. In our implementation, we used a Keyence BZ-X710 inverted fluorescence phase contrast microscope, for which we provide open-source image stitching code - runs in any computer, independent of the software provided by the manufacturer.

Warning

If you use a different microscope, please refer to the documentation of your microscope for how to stitch tile-scans into a single image. The Open-ST pipeline expects one single file in either tiff, jpeg or png formats.

Stitching with the script

When using our same setup, we provide a script that automatically handles the imaging data generated by the Keyence microscope and leverages the Grid/Collection stitching plugin included in Fiji to create a single, composite image of the tile-scan. Open a terminal, and run the following command:

openst image_stitch \
     --microscope='keyence' \
     --imagej-bin=<path_to_fiji_or_imagej> \
     --tiles-dir=<path_to_tiles> \
     --tiles-prefix=<to_read> \
     --tmp-dir=<tmp_dir> \
     --output-image=<output_image>

Make sure to replace the placeholders (<...>). For instance, <path_to_fiji_or_imagej> is the path where the Fiji executable is; <path_to_tiles> is the folder where the microscope has saved the individual image files with the tile scan and the .bcl file with metadata; <to_read> is the prefix that is common to all the tile image files, and <tmp_dir> is a path with write permission where the images will be temporarily moved. Finally, <output_image> is the full path and file name that will be given to the stitched image (must be a writeable folder).

Question

If you don't know how to specify the <path_to_fiji_or_imagej>, please follow the official instructions provided for Running Headless

(Optional) Addressing image irregularities

Most of the times, large images have uneven illumination, focus or noise. This can be challenging when doing downstream processing on these images, like segmentation, feature extraction or quantification (i.e., on fluorescence images). There is a plethora of methods to address these issues (e.g., Flatfield Correction from BigStitcher, or CARE, to name some). This might be highly dependent on your microscope, imaging settings, sample type, sample width... Always look at your images so you can take an informed decision.

In our publication, we leveraged a CUT model that allowed us to homogeneize the style of the whole tile-scan - that is, reduce possible biases in illumination, noise and focus - across the entire tile-scan. You can run this by running the following command on the stitched image.

openst image_preprocess \
     --input=<path_to_input_image> \
     --CUT \
     --CUT-model=<path_to_model> \
     --output=<path_to_output>

Make sure to replace the placeholders (<...>). For instance, <path_to_input_image> is the full path and file name of the previously stitched image; <path_to_model> is filename our pre-trained CUT model, and <output_image> is the path to a folder (writeable) and desired filename for the output image.

Expected output

After running the stitching (and optionally correction algorithms), you will have a single image file per sample. This, together with spatial transcriptomics data from the previous step, will be used in the following to align both modalities, and eventually obtain a file that can be used for the downstream spatial, single-cell analysis.