4. QuPath for analysing ENS
Qupath is a great software used in digital pathology especially for large whole slide images. It is well designed, has an intuitive and responsive interface with great support. QuPath deals with analysing multiplex data particularly well. If you would like to analyze large tilescans, using QuPath may be a better choice.
-
Follow the instructions here to download and install QuPath: https://qupath.readthedocs.io/en/stable/docs/intro/installation.html
-
For GAT, we need the StarDist extension, so please install that from here: https://qupath.readthedocs.io/en/stable/docs/advanced/stardist.html#getting-the-stardist-extension
-
In addition to GAT, we will need the GAT model and a QuPath script. Download them from here. To download single files from github, follow this instruction video here. If you'd like to download multiple files, go to next section in same video.
Tutorial videos for QuPath are available here:
You can use Cellpose within QuPath for segmenting cells now. For more details, check here: https://github.com/BIOP/qupath-extension-cellpose It does require configuring a Python environment with Cellpose and installing a QuPath extension. This enables access to state-of-the-art segmentation models from within QuPath.
Tutorial to be added soon.
The workflow for segmenting neurons can be used to get an outline of each enteric neuron. This is quite similar to many workflows in QuPath where you segment the nuclei using the DAPI channel first and then perform downstream analysis to characterise each cell. We can use a similar workflow if we are looking at multiple markers or neuronal subtypes. An example video is available in the tutorial lin k above. However, a general workflow will look like:
- Segment the neurons using the enteric neuron Stardist model
- Perform cell classification using:
- Thresholding Approach
- Machine learning (ML) classifier approach
Both these approaches are mentioned in detail here:
https://qupath.readthedocs.io/en/stable/docs/tutorials/multiplex_analysis.html
You can perform the Cell detection part of the tutorial by segmenting the neurons with the stardist model approach. This can be followed up by the thresholding or ML approach to classify cells based on different channels/other criteria.
This section will keep being updated...