scMultipleX provides functions for image processing workflows for high-content, multiplexed, and volumetric microscopy datasets.
The package supports workflows for segmentation post-processing, object linking across multiplexing rounds, registration, intensity correction, feature extraction, mesh generation, and 3D shape analysis.
scMultipleX functions are wrapped as Fractal tasks based on the OME-Zarr image format. scMultipleX tasks enable users to build reproducible image-processing workflows from modular tasks and execute them locally or on compute clusters.
The tasks provided by scMultipleX are located under:
src/scmultiplex/fractal/
and can be collected and executed within a Fractal deployment alongside tasks from the Fractal ecosystem.
Tasks for preparing and cleaning label images:
- Build label images
- Expand labels
- Fuse touching labels
- Clean up 3D cell segmentations
- Segment images by intensity threshold
- Convert 3D images to maximum-intensity projections
Tasks for linking related objects across segmentations or imaging rounds:
- Calculate object linking
- Calculate linking consensus
- Relabel objects using linking consensus
- Link parent and child objects, such as nuclei within cells or organoids
Tasks for aligning multiplexed imaging rounds:
- Calculate Warpfield registration for 3D pixel-based registration
- Apply Warpfield registration
- Calculate and shift images by rigid transformation (translation and rotation)
- Post-registration cleanup
- Detect clipped ROIs across rounds
Tasks for correcting Z-dependent illumination effects:
- Calculate z-illumination correction
- Apply z-illumination correction
Tasks for extracting quantitative measurements from segmented objects:
- Morphology measurements
- Intensity measurements
- Pixel-threshold measurements
- Measurements within masked parent ROIs
- Regionprops-based 2D and 3D feature tables
Tasks for generating and analyzing 3D surface meshes:
- Generate multiscale surface meshes
- Measure mesh-based features
- Calculate spherical harmonics from label images
- Annotate child meshes
- Annotate meshes using child-object features
Helper tasks for selecting images and defining parallelization lists:
- Select single imaging rounds
- Select multiple imaging rounds
- Select multiplexing pairs
- Select reference rounds
- Select illumination-correction rounds
git clone https://github.com/fmi-basel/gliberal-scMultipleX.git
cd gliberal-scMultipleX
pip install -e .pip install -e ".[spherical-harmonics, warpfield]"BSD 3-Clause License.
Developed in the Liberali Lab at the Friedrich Miescher Institute for Biomedical Research (FMI), Basel, Switzerland by @nrepina, with development support from @jluethi, @lorenzocerrone, @enricotagliavini, and @tibuch.
Unless otherwise stated in each individual module, all scMultipleX components are released according to a BSD 3-Clause License, and Copyright is with Friedrich Miescher Institute for Biomedical Research.
Point-cloud based multiplexed linking is built on PlatyMatch (DOI) by Manan Lalit.
Warpfield registration is based on Warpfield (DOI).
Spherical harmonic computation is based on aisc-shparam.