mCRC Manuscript Scripts

This repository contains all R scripts and analysis pipelines used for the manuscript: "Spatial Zonation of Tumor Cell States, Stromal-Immune Networks, and Growth Patterns in Metastatic Colorectal Cancer". Scripts are organized by analysis category and figure generation.

Folder Structure

• A_Data_preprocessing/ - Data preprocessing pipelines
  • A_snMultiome/ - Single-nucleus multiome (RNA + ATAC) processing
  • B_Xenium/ - Xenium spatial transcriptomics processing
  • C_CellChat/ - Cell-cell communication analysis
  • D_AUCell/ - Gene set enrichment analysis
• B_Data_collection_visualization/ - Data collection overview figures
• C_Somatic_mutation_analysis/ - Mutation analysis and visualization
• D_Bulk_RNAseq/ - Bulk RNA-seq analysis
• E_NB_and_Functional_Zone/ - Neighborhood and functional zone analysis
• F_Tumor_subtypes/ - Tumor subtype analysis
• G_TMR_PSI_interaction/ - Tumor Margin Region and Peri-Stromal Interface analysis
• H_Organ_specific_adaption/ - Organ-specific adaptation analysis

Requirements

• R: 4.4.3 (via seurat5_env conda environment)
• Python: 3.13.3 (R analysis), 3.13.7 (3D analysis), 3.11.6 (Banksy clustering), 3.10.19 (morphological annotation)
• Key packages: Seurat v5.3.0, tidyverse, ComplexHeatmap, AUCell, CellChat
• System: Linux (tested on RHEL 7), 30GB+ RAM recommended

Quick Start

# R analysis (most scripts)
conda env create -f envs/seurat5_env.yml
conda activate seurat5_env

# Morphological annotation
conda env create -f envs/morph_env.yml
conda activate morph_env
pip install git+https://github.com/ding-lab/morph.git

# 3D reconstruction
conda env create -f envs/3d-analysis_env.yml
conda activate 3d-analysis

# Banksy clustering (Xenium)
conda env create -f envs/banksy_env.yml
conda activate banksy

See envs/README.md for build times and detailed setup instructions. For package documentation, see PACKAGE_DOCUMENTATION.md.

Bioinformatics Tools

Many bioinformatics tools were used in the course of this work. All tools written and/or published by the authors are freely available at our public GitHub repository (https://github.com/ding-lab/), including:

• somaticwrapper - Variant calling pipeline: https://github.com/ding-lab/somaticwrapper
• 10Xmapping - Code for mutation mapping from bulk to single cells: https://github.com/ding-lab/10Xmapping
• ffpefiltering - Pipeline for FFPE WES filtering: https://github.com/ding-lab/ffpefiltering
• pecgs-pipeline - Bulk RNA-seq alignment and transcript counting: https://github.com/ding-lab/pecgs-pipeline
• Morph - Spatial transcriptomics toolset for tumor boundary detection: https://github.com/ding-lab/morph

Usage

Scripts are named with figure identifiers (F#X_ for main figures, S#X_ for supplementary). Set working directory to script location and source:

setwd('/path/to/mCRC_Manuscript_Script/[folder_name]')
source('script_name.R')

Note: Scripts require access to specific server file paths and data directories. Many depend on custom functions in /diskmnt/Users2/epeng/Projects/mCRC/scripts/.

---

Manuscript: Spatial Zonation of Tumor Cell States, Stromal-Immune Networks, and Growth Patterns in Metastatic Colorectal Cancer
Status: Under Review