Hi there! We are excited to introduce BoReMi, a Python-based registration tool powered by the Bokeh visualization library. BoReMi simplifies the manual alignment of spatio-molecular data to microscopy images, offering a user-friendly interface within Jupyter Notebooks.
A big thank you to the Bokeh team for developing an incredibly powerful visualization library!
Why BoReMi?
Spatio-molecular data and microscopy images provide complementary information, essential to study structure and function of spatially organised multicellular systems such as healthy or diseased tissues. However, aligning them is challenging due to differences in scale, orientation, and coordinate space.
While automated tools are evolving, manual alignment is still crucial for:
- Improving ambiguous or suboptimal automated results.
- Solving artefacts using human insight.
- Generating ground-truth datasets for benchmarking and training automated tools.
BoReMi addresses these needs by integrating manual alignment directly into Python-based workflows, eliminating tedious workarounds required by tools like ImageJ or GIMP.
Target Audience
BoReMi is designed for researchers working in spatial biology, but researchers from other fields may also find its implementation helpful. Discussions like these have inspired and guided us, and we hope BoReMi does the same for others.
Introducing BoReMi
BoReMi (available on GitHub with an interactive Binder demo, supporting local and cluster-based setups) is an interactive, Jupyter-embedded tool designed for registering spatio-molecular data with microscopy images. It supports:
- Linear (rigid) manipulations: scaling, rotation, translation, flipping.
- Non-linear (elastic) adjustments for complex registration needs.
- Seamless integration within Jupyter Notebooks for efficient workflows.
Note: Due to uploading limitations, we have included an animated GIF as a demonstration. For a hands-on experience, please have a look at the Binder in our GitHub repository for an interactive demo of BoReMi!
Key Features
-
Interactive Visualization
Intuitive Bokeh-powered interface for precise alignment, supporting both linear and non-linear transformations. -
Multi-Dataset Handling
Manage multiple sets of spatio-molecular data simultaneously for comprehensive analysis. -
Convenience Features
- Zoom and pan for precise adjustments.
- Undo, redo, and reset functionality for flexibility.
- Flexible input and output options for seamless data handling.
- Image gallery to preview associated images.
- Cluster selection for enhanced pattern recognition.
- Efficient Performance
Built-in downsampling ensures smooth interactivity, even for large datasets.
Applications and Future Directions
Manual alignment tools like BoReMi serve as the foundation for generating high-quality ground-truth datasets which are essential for benchmarking and training automated alignment tools. Future improvements include faster data loading, continuous value display of data points, and expanded non-linear adjustment capabilities.
We hope BoReMi proves helpful to the community and warmly welcome any feedback or ideas you might have for improving it!
Publication
BoReMi is featured in PLOS Computational Biology. Full paper here!