The APx Fractal Task Collection is mainainted by Apricot Therapeutics AG, Switzerland. This is a collection of tasks intended to be used in combination with the Fractal Analytics Platform maintained by the BioVisionCenter Zurich (co-founded by the Friedrich Miescher Institute and the University of Zurich). The tasks in this collection are focused on extending Fractal's capabilities of processing 2D image data, with a special focus on multiplexed 2D image data. Most tasks work with 3D image data, but they have not specifically been developed for this scenario.
You can install the package locally with:
pip install git+https://github.com/Apricot-Therapeutics/APx_fractal_task_collection.gitFor instructions on how to run tasks, please refer to the official Fractal Tasks Core Documentation
Please note that all tasks pass basic tests based on 2D and 3D OME-ZARR files. However, in the 3D case, the resulting output has not been extensively checked and might not make sense (as most tasks have been developed for the 2D use-case). If you are using the tasks for 3D data and encounter any weird behaviour, please open an issue.
| Task | Description | 2D Tests Passing | 2D Output validated | 3D Tests Passing | 3D Output validated |
|---|---|---|---|---|---|
| Convert IC6000 to OME-Zarr | Converts output images from IN Cell Analyzer 6000 (GE Healthcare) to OME-Zarr. | ☑️ | ☑️ | ✖️ | ✖️ |
| Calculate BaSiCPy Illumination Models | Calculates illumination correction model based on BaSiCPy for each available channel. | ☑️ | ☑️ | ☑️ | ✖️ |
| Apply BaSiCPy Illumination Model | Applies BaSiCPy illumination models to a OME-Zarr file. Use after "Calculate BaSiCPy Illumination Models" | ☑️ | ☑️ | ☑️ | ✖️ |
| Correct Chromatic Shift | Corrects chromatic shift in OME-Zarr file per wavelength id. Requires reference images (for example fluorescent beads) | ☑️ | ☑️ | ☑️ | ✖️ |
| Stitch FOVs with Overlap | Stitches FOVs that were imaged with overlap, using ASHLAR. | ☑️ | ☑️ | ✖️ | ✖️ |
| Segment Secondary Object | Segments secondary objects in images. Requires a label image that provides seeds and an intensity image. | ☑️ | ☑️ | ☑️ | ✖️ |
| Detect Blob Centroids | Performs blob detection using scikit-image | ☑️ | ☑️ | ☑️ | ✖️ |
| Clip Label Image | Clips a label image with a secondary label image. For example, this can be used to clip cell segmentations with nuclear segmentations to receive the cytoplasm. | ☑️ | ☑️ | ☑️ | ✖️ |
| Mask Label Image | Applies a mask to a label image based on a secondary label image. | ☑️ | ☑️ | ☑️ | ✖️ |
| Convert Channel to Label | Utility task to convert a channel from a OME-Zarr file to a label image. Can be used to import an external label image into Fractal without creating a new task. | ☑️ | ☑️ | ☑️ | ✖️ |
| Filter Label by Size | Filters a label image by size and removes objects larger/smaller than a given threshold. | ☑️ | ☑️ | ☑️ | ✖️ |
| Measure Features | Measures features for a given label image. Currently, four feature sets are available: intensity featues, morphology features, population context features and texture features (Haralick and Law's Texture Energy). | ☑️ | ☑️ | ☑️ | ✖️ |
| Label Assignment by Overlap | Assigns child labels to parent labels by their overlap. Relationship is saved in the observations of the feature table. | ☑️ | ☑️ | ☑️ | ✖️ |
| Aggregate Feature Tables | Aggregates/Concatenates feature tables from all multiplexing acquisitions. Can be saved either on the well level or in the first aqcuisition. | ☑️ | ☑️ | ☑️ | ✖️ |
| Multiplexed Pixel Clustering | Applies multiplexed pixel clustering to selected images. | ☑️ | ☑️ | ☑️ | ✖️ |