User guide

• ComSeg tutorial
  • 1) dataset initialization
    • Computation of the co-expression correlation at the dataset scale
    • The estimated co-expression correlation matrix between genes can be vizualised as folow
  • 2) Create a graph of RNA molecule nodes and apply graph patitioning
  • 3) In situ clustering
  • 4) final RNA-nuclei association
• Minimal example : run ComSeg with a config dictionary
  • Run ComSeg with configuration dictionary.
  • Plot result
  • Comprensive description of the configuration dictionnary
    • dataset initialisation
    • Co-expression computation
    • knn graph generation
    • in-situ clustering
    • rna association to centroid
    • final result
    • final result
• ComSeg on large scale data with SOPA / SpatialData
  • ComSeg is now integrated in SOPA : https://gustaveroussy.github.io/sopa/tutorials/comseg/. This tutorial is depreciated
    • Load example data
    • Segmentation of nuclei with Cellpose
    • Compute the patches for ComSeg
    • Compute centroid for ComSeg
    • Run ComSeg on each patch
    • Aggregate results
    • Plot results