Loom, is a visual computing system that helps researchers examine pseudo-temporal cellular transitions, spatial enrichment, and gene-expression dynamics within local microenvironments while enabling comparisons across samples and regions of interest.It integrates commonly used tools and data structures, including Scanpy, AnnData, and Squidpy, while providing an intuitive way to select regions of interest and explore how cellular activity changes across space and time.

One common challenge in spatial transcriptomics is that analysis workflows are often fragmented across multiple tools. Researchers may need to move between different packages for data processing, spatial analysis, visualization, and region-specific exploration.

Primary Use Cases:

• Cell State Tracking: Analyze how specific cell types change their state or behavior.
• Spatial Microenvironments: Track cellular changes relative to physical distances from key tissue structures.
• Feature Proximity: Study cell behavior near blood vessels or tumor boundaries.

Loom aims to bring these steps together in a more unified workflow.

Some features that stood out to me:

• Direct support for publicly available 10x Genomics spatial transcriptomics datasets
• A one-command workflow for downloading and processing supported datasets
• No need to manually perform each preprocessing step
• Transparent documentation explaining the logic behind every processing stage
• Integrated Scanpy, AnnData, and Squidpy functionality
• Region-of-interest exploration across spatial and temporal dimensions
• Public example datasets and reproducible workflows
• One-command Docker installation for easier setup

The GitHub repository includes example data, so users can install the environment and test the workflow without first preparing their own dataset.

What I find especially valuable is that the project does not treat the processing pipeline as a black box. The repository explains the reasoning and implementation behind each step, which makes the workflow easier to understand, verify, and adapt.

This could be useful for researchers studying tissue development, disease progression, tumor microenvironments, cellular interactions, or other spatially dynamic biological processes.

GitHub: https://github.com/ScheWann/Loom

Paper: https://arxiv.org/abs/2607.22505

For people working with spatial transcriptomics:

• Would a unified workflow like this be useful in your research?
• Which additional datasets or integrations would you want it to support?
• How does this compare with your current Scanpy or Squidpy workflow?

Any feedback, advice, objection .. whatever is highly appreciated in advance! you can also reach us directly at szhao69@uic.edu!