This repository contains a pipeline for single-cell-level analysis of large, high-resolution, multiplexed whole-tissue images. The workflow includes image processing steps implemented in Jupyter notebooks, Cellpose-based segmentation, CellProfiler pipeline for feature extraction, and R-code for data handling and analysis.

Large high-resolution images are tiled and processed for segmentation. Cellpose segmentation is performed at full resolution for tiled images, after which images are downscaled to reduce data size. Single-cell properties, including marker expression and spatial coordinates, are measured in CellProfiler. Finally, data is loaded into R and mapped back to the whole tissue scale.

1. Install conda and git if they are not already installed.

2. Open conda prompt and clone the repository :

git clone https://github.com/jonasiaho/WholeTumorSeg.git

3. Change directory to WholeTumorSeg cd WholeTumorSeg

4. Create new conda environment with env.yml :

conda env create -f env.yml

5. Check that "wholetumorseg" environment was succesfully created with conda env list

See documentation

We provide example data set for testing the pipeline.