Spatial Cellular Neighborhood

Summarize cell-type neighborhoods from annotated spatial immune microenvironment data

What it does

This workflow focuses on cellular-neighborhood analysis in the immune tumor microenvironment rather than on primary spatial preprocessing. The committed README describes a Python notebook-based method for identifying major components of association between cell types and cellular neighborhoods from an already annotated table.

When to use it

Use this workflow when the upstream spatial or imaging data have already been annotated with cell types and neighborhood labels, and the next question is how neighborhoods relate to cell-type composition across samples or patient groups. It is a focused downstream analysis branch rather than a full spatial transcriptomics pipeline, and it assumes the neighborhood annotation step has already happened elsewhere.

Prerequisites

  • Source folder: Spatial_Cellular_neighborhood
  • Main documentation: readme.md
  • Main implementation note in the README: download the Python notebook separately and run it in Jupyter
  • Expected input table columns:
    • patients
    • optional groups
    • neighborhood10
    • ClusterName
  • Sample data linked from the README:
    • CRC_clusters_neighborhoods_markers.csv

Steps

Prepare an annotated CSV with the required columns

The README makes the input contract explicit: the CSV must already contain sample identifiers, neighborhood labels, and cell-type annotations.

The required columns listed in the committed source are:

  • patients
  • optional groups
  • neighborhood10
  • ClusterName

Open the notebook and point path_to_data at the CSV file

The committed instructions say to download the Python code, open it in Jupyter, and change path_to_data to point to the CSV file.

Run the notebook sequentially to summarize neighborhood associations

The intended outcome is to identify the main associations between cell types and cellular neighborhoods in the tumor microenvironment.

This page stays concise because the repository only commits the README, not the notebook itself. The source materials make the operational sequence clear even if they do not expose the intermediate code cells on the site.

Use the sample dataset and cited paper as context, not as in-repo assets

The README links to an external sample file, CRC_clusters_neighborhoods_markers.csv, and to the original publication for methodological detail. Those references are useful for understanding the expected input shape and the intended analysis context, but they are external resources rather than committed local assets.

Gotchas / notes

  • This folder is thin: the README is committed, but the Python notebook itself is not included in the repository.
  • The method assumes neighborhood annotations already exist upstream; it does not perform segmentation or primary spatial preprocessing.
  • The sample dataset is external rather than bundled in the repo.
  • The README points to the original publication for methodological detail, which means this site page necessarily stays concise and operational.

📄 View source on GitHub