T. Truc Bui, Max-Malte Hansen, Jan P. Hummel, Julius J. Stein, Korbinian Traeuble

Technical University of Munich & Helmholtz Munich

We mapped endothelial cell subtypes using Visium spatial transcriptomics and Cell2Location, with a focus on pro-angiogenic endothelial cells. By integrating single-cell RNA sequencing (scRNA-seq), we enhanced spatial resolution to uncover their distribution, function, and role in angiogenesis within tissue microenvironments.

First, create a virtual environment using venv:

python -m venv venv

• Windows (cmd/PowerShell):

  venv\Scripts\activate

• Mac/Linux: (currently no model training with mps)

  source venv/bin/activate

Once the virtual environment is activated, install the required dependencies:

pip install -r requirements.txt

In the root directory of the project, create a .env file and copy-paste the following contents into it:

# Example .env file
PATH_TO_SCATLAS=/PATH/TO/PLAQUE/ATLAS.h5ad
PATH_TO_VISIUM=/PATH/TO/VISUM/SLIDES.h5ad
PATH_TO_ATLAS_MODEL2WO=./
PATH_TO_MAP_MODEL2WO_N5=./
PATH_TO_MAP_MODEL2WO_N7=./
PATH_TO_MAP_MODEL2WO_N10=./

Ensure everything is installed correctly by running:

python -m pip check

Script to create a diff of 3 plots with different hyperparameters [specifically cells per location]. Make sure, the plots have the same dimensions.

To run the script:

python3 compare_plots.py plot1.png plot2.png plot3.png [output.png]

plot1: Path to the first plot image file.
plot2: Path to the second plot image file.
plot3: Path to the third plot image file.
--output: Path to save the output comparison image (optional)

python3 compare_plots.py n5.png n7.png n10.png -o diff_ncells_5_7_10.png

Visium Data: Bleckwehl et al. (2024)
Cell2location: Kleshchevnikov et al. (2022)
Single Cell Atlas: Traeuble et al. (2024)