Xenflow

Xenium spatial transcriptomics analysis pipeline built with Nextflow and Singularity/Apptainer. The pipeline uses cv2 to automatically identify tissue contours and bounding boxes; it is designed to analyse tissue microarrays and multi-sample slides.

Quick start

1. Build a container sopa.sif from the definition file from this repository.
2. Export CONTAINERDIR and PROJECTDIR variables. The former is used to find the container, and the latter to mount the data to singularity during run.
3. Adjust resources in resources.config for pipeline steps.
4. Create run01.config in config directory from run.template to configure step parameters.
5. Create environment with nextflow installation using make env.
6. Run the pipeline using make run.

Key steps

The pipeline has 4 steps starting from raw xenium output to identification of gene programs using cNMF.

• CONVERT_XENIUM converts raw data to spatialdata-formatted zarr archive for downstream processing.
• DETECT_TISSUE automatically identifies tissue contours and bouding boxes to split multi-sample slides into separate sample objects for independent analysis. Especially helpful when working with tissue microarrays.
• SPLIT_SAMPLES creates one AnnData h5ad archive per sample based on the tissue contours.
• IDENTIFY_PROGRAMS runs cNMF for each sample. Note that it doesn't automatically select the best number of programs in this pipeline leaving this task for user.