Single-Cell RNA-seq Analysis Pipeline

🧬 A comprehensive end-to-end pipeline for single-cell RNA sequencing data analysis, from quality control to cell-cell communication inference.

🔬 Overview

This repository provides a complete workflow for analyzing single-cell RNA-seq data using state-of-the-art computational methods. The pipeline integrates both R/Seurat and Python/scVI frameworks to deliver robust results from raw data to biological insights.

Key Features:

• Quality control and doublet detection
• Multiple normalization strategies (log1p, SCTransform)
• Data integration using Seurat V5/Harmony and scVI
• Automated and manual cell type annotation
• Differential abundance testing with MILOR
• Multiple cell-cell communication methods
• Differential expression analysis
• Comprehensive visualization tools

📁 Pipeline Structure

1. Quality Control & Preprocessing

• 1.QC_DoubletF_Log1p.qmd - QC pipeline with doublet filtering and log1p normalization
• 1.QC_DoubletF_SCT.qmd - Alternative QC using SCTransform normalization and doubletFinder
• basic_analysis_steps_MISC.Rmd - Additional QC utilities and helper functions

2. Data Integration

• 2.Data intergration.Rmd - Batch correction and sample integration (Seurat V5)
• Single_cell_scVI.ipynb - Deep learning-based integration using scVI
• Subclustering_SCVI.ipynb - Subclustering analysis with scVI
• change_scvi_continuous.ipynb - Handling continuous covariates in scVI

3. Cell Type Annotation

• 3.Annotation.Rmd - Automated and manual cell type annotation
• High_annotation_multicontrast.Rmd - High-resolution annotation across conditions
• High_level_multicontrast.Rmd - Broad cell type classification
• Low_annotation_multiniche_multicontrast.Rmd - Fine-grained annotation analysis

4. Differential Abundance Analysis

• 4. MILOR_dif_abud.Rmd - MILOR-based differential abundance testing

5. Subclustering & Advanced Analysis

• 5.Subclustering.Rmd - Detailed subclustering of cell populations
• Subclustering_SCVI.ipynb - Python-based subclustering with scVI

6. Cell-Cell Communication Analysis

• 5.cc_interactions_niche.Rmd - Niche-based cell communication analysis
• 5.cellchat.Rmd - CellChat communication inference
• 6.cc_interactions_niche.Rmd - Extended niche interaction analysis
• 6.cellchat.Rmd - Advanced CellChat workflows
• 6.cellphonedb.Rmd - CellPhoneDB communication analysis
• 6.Multinichenet.Rmd - MultiNicheNet analysis
• 6.1.Multinichenet_output_analysis.Rmd - MultiNicheNet results interpretation

7. Differential Expression Analysis

• 7.DEG_conditions_subtypes.Rmd - Differential gene expression across conditions and cell types

8. Metabolic Analysis

• mebocost_rasV.ipynb - Metabolic cost analysis using MEBOCOST

9. Visualization & Utilities

• tSNE.Rmd - t-SNE dimensionality reduction and plotting
• UMAP_color_Change.Rmd - UMAP visualization customization
• relative_percentages_plots.Rmd - Cell proportion visualization
• object_format_convert.Rmd - Format conversion utilities

🚀 Quick Start

Prerequisites

# R packages (Seurat V5 required)
install.packages("Seurat") # Version 5.0+
install.packages(c("SingleCellExperiment", "scater"))
BiocManager::install(c("miloR", "CellChat", "MultiNicheNet"))

# Python packages
pip install scvi-tools scanpy pandas numpy matplotlib seaborn

Basic Usage

1. Start with QC: Run 1.QC_DoubletF_Log1p.qmd or 1.QC_DoubletF_SCT.qmd
2. Integrate data: Use 2.Data intergration.Rmd or Single_cell_scVI.ipynb
3. Annotate cells: Apply 3.Annotation.Rmd
4. Analyze communications: Choose from CellChat, CellPhoneDB, or MultiNicheNet scripts
5. Find DEGs: Run 7.DEG_conditions_subtypes.Rmd

📊 Methods Implemented

Quality Control:

• Doublet detection and filtering
• Mitochondrial gene filtering
• Low-quality cell removal

Normalization:

• Log1p normalization
• SCTransform
• scVI normalization

Integration Methods:

• Seurat V5 CCA/RPCA/Harmony
• scVI deep learning integration

Cell-Cell Communication:

• CellChat
• CellPhoneDB
• MultiNicheNet
• Custom niche analysis
• Mebocost

Differential Analysis:

• MILOR (differential abundance)
• Seurat V5 FindMarkers/FindAllMarkers
• edgeR/DESeq2 integration

📈 Expected Outputs

• Quality control reports and plots
• Integrated single-cell object
• Cell type annotations
• UMAP/t-SNE visualizations
• Cell-cell communication networks
• Differential expression results
• Abundance analysis results

🛠️ File Formats

• Input: 10X Genomics H5/MTX, CSV, H5AD, RDS
• Output: Seurat objects (RDS), AnnData (H5AD), CSV results, HTML reports

---

Note: This pipeline was developed for a private murine dataset. Adjust parameters as needed for your dataset.