Exploring spatial omics made simple with KaroSpace

If you work with spatial transcriptomics or multimodal spatial datasets, you’ve likely faced the same bottleneck: moving from processed data to actual biological insight. Visualization across samples, conditions, and modalities is often fragmented, slow, or requires custom tools.

KaroSpace addresses this directly by turning your dataset into an interactive exploration environment, where you can move continuously between cells, tissue context, and samples.

Recently released as a preprint, KaroSpace is a rapid-access, cell-centric framework designed for interactive exploration of multi-sample and multi-modal spatial omics data, while remaining agnostic to upstream pipelines.

What makes KaroSpace particularly useful?

• Interactive browser for spatial datasets

Generate a light weight ready-to-use HTML to explore your data (locally or shared) without complex deployment.

• Cell-centric exploration across samples

Navigate seamlessly between cells, regions, and conditions, enabling direct comparison of spatial patterns across datasets.

• Multi-sample and multi-modality support

Designed to handle the growing complexity of spatial omics, including integration across technologies and experiments.

• Pipeline-agnostic design

Works independently of upstream preprocessing, making it easy to integrate into existing spatial workflows.

• Fast setup with open-source tooling

The associated ecosystem (including builder tools and GitHub resources) allows rapid generation of custom viewers for your own datasets.

Why this matters?

Spatial omics is fundamentally about context, understanding how cells are organized and interact within tissue architecture. Instead of exporting static plots, KaroSpace allows you to interact directly with the dataset, making exploration intuitive, scalable, and shareable, which is critical as datasets become larger and more complex. It enables faster iteration between hypothesis and validation, especially in projects involving multiple sections, timepoints, or conditions.

Getting started with KaroSpace (practical setup)

1.Prepare your data

• Start from a processed dataset (typically AnnData or similar structure)
• Ensure you have :
  • Cell-level metadata (clusters, annotations, conditions)
  • Spatial coordinates
  • Expression matrix (and optionally additional modalities)

KaroSpace is not a preprocessing tool, it expects analysis-ready data.

2.Create your KaroSpace

• Download the GUI builder
• Set input .h5ad and output directory
• Click Inspect H5AD
• Choose groupby/color settings (or a preset)
• Select your genes of interest
• Click Export

or via command line

3.Launch the interface

• Once built, double click on your HTML file.

You now have a fully interactive spatial exploration interface, typically only a few megabytes in size, making it easy to share via email or discussion channels.

Upcoming

A sidecar solution is currently under development to convert fully processed datasets and complete gene matrices into KaroSpace punch cards, which can then be directly uploaded to a web-hosted visualizer (KaroSpace Package Loader on the website).

Resources

• Preprint
• Website
• GitHub

If you are working with technologies like Xenium, Visium, MERFISH, or integrated spatial datasets, this is worth testing. The ability to quickly spin up an interactive, shareable spatial viewer can significantly streamline both analysis and collaboration.

Curious to hear any feedback or comment or improve this multi-sample spatial visualization approach even further.