xenium-liver-cancer-data-analysis

This repository contains preprocessing workflows and reference-based cell typing pipelines for analyzing human liver data from the Xenium platform by 10x Genomics. The dataset includes both healthy and diseased (cancerous) liver tissue samples, enabling spatial transcriptomic comparisons between normal and pathological liver states.

📊 Data

Source: 10x Genomics – Human Liver Data (Xenium)

📥 Data Access To download and use the Xenium dataset:

Visit the dataset page

Download the dataset ZIP file(s) (Liver cancer -> xenium_data/diseased and Non-diseased -> xenium_data/non_diseased)

Extract and place them under xenium_data/ directory in this repo

📌 Project Goals

• Preprocess raw Xenium spatial transcriptomics data.
• Perform quality control and normalization steps.
• Conduct reference-based cell type annotation using external single-cell RNA-seq datasets.
• Explore spatial distribution of cell types across healthy and cancerous liver tissues.
• Lay the foundation for further downstream analyses such as differential expression, spatial domain discovery, and tumor microenvironment characterization.
• 📈 Preliminary Results: Analysis results and figures will be added as the project progresses.

🗂️ Repository Structure

xenium_liver_cancer/
|
├── xenium_data/                      # Input Xenium datasets (healthy & diseased)
    |-diseased
    |-non_diseased
├── scripts/                   # Modular Python/R scripts for pipeline steps
├── results/                   # Output figures, cell type labels, etc.
├── environment.yml            # Conda environment YAMLs
├── README.md                  # Project overview
└── ....                       # python classes and helper functions

🛠️ Tools and Technologies

- Python: Main programming language for data preprocessing and analysis.
- Scanpy: For single-cell and spatial transcriptomics analysis.
- Anndata: Data structure for annotated data matrices.
- Zarr: Efficient storage of large, chunked, compressed, N-dimensional arrays.
- R (optional): For additional analysis and visualization (e.g., Seurat).
- Jupyter Notebooks: Interactive analysis and visualization.
- Conda: Environment and dependency management.

📌 To Do

✅ Setup project structure

Preprocess Xenium liver cancer and healthy data

Perform reference-based cell typing

Visualize spatial cell distributions

Add documentation and analysis figures

📚 Documentations

• FeatureMatrixLoader
• CellStatloader

🙋 Acknowledgments

10x Genomics for providing the Xenium platform and datasets.

The broader open-source community for tools used in spatial transcriptomics.