Odon: an ultra-fast viewer for spatial proteomics

Area of Science:

• Computational biology
• Bioinformatics
• Data visualization

Background:

• Multiplexed spatial imaging generates large, high-dimensional datasets.
• Efficient visualization is crucial for identifying staining artifacts and analyzing data at scale.

Purpose of the Study:

• To develop a high-performance desktop viewer for rapid, interactive exploration of multiplex imaging data.
• To enable efficient analysis of spatial proteomics and transcriptomics datasets on standard hardware.

Main Methods:

• Developed Odon, a native Rust desktop viewer utilizing the OME-Zarr format.
• Implemented viewport-driven tile loading and GPU-based compositing for optimized rendering.
• Supported GeoJSON, GeoParquet, SpatialData, Xenium containers, and TIFF formats.

Main Results:

• Odon demonstrated lower peak memory use and faster warm-start image loading compared to napari and QuPath.
• Achieved smooth rendering and interaction with over 1,000,000 segmented cells.
• Integrated visual analytics features like live thresholding and mosaic mode for cohort studies.

Conclusions:

• Odon provides a scalable, high-performance platform for visualizing and analyzing large multiplex imaging datasets.
• Its efficient design facilitates rapid exploration and artifact detection on standard laptops.
• Enables advanced visual analytics for cohort and tissue microarray studies.