Welcome¶
Under development
This tool is still under active development.
Spellmatch is a command-line tool for spatial cell representation matching in segmented 2D images. Its main use case is multi-panel imaging, where consecutive tissue slices are imaged with partially overlapping panels, e.g. to enhance multiplexity.
Spellmatch matches spatial cell representations
Unlike other approaches, spellmatch is not meant for matching similar cells, but matches different representations of the same cell. Importantly, spellmatch only includes algorithms for spatial cell matching, i.e. algorithms that consider the spatial distributions of cells in tissue.
Overview¶
Spellmatch provides the following functionality:
- Interactive cell matching for ground truth generation and initial image alignment
- Automatic feature/intensity-based image registration for refining image alignments
- Point set registration for matching cells based on their spatial distribution alone
- Network alignment for matching cells based on their topology and attributes (e.g. intensities)
- Extraction of directed and/or undirected cell-cell assignments from cell alignment scores
Notably, spellmatch ships with its own network alignment algorithm, which leverages the full information present in multichannel images to obtain cell alignment scores. For an overview of all matching algorithms implemented in spellmatch, please refer to Algorithms.
For developers, spellmatch can also be used as a Python package, e.g. to extend spellmatch or to integrate it into custom pipelines. Please refer to Reference for a detailed API documentation
Getting started¶
Getting help¶
A manuscript is in preparation.
For conceptual questions, please do not hesitate to reach out.
For technical issues, please refer to the spellmatch issue tracker.
Citing spellmatch¶
A manuscript is in preparation.
If you are using spellmatch in your work, please reach out to inquire how to best cite the tool.