scRNA-seq Analysis Tutorial: From Raw Counts to Cell Communication

Dataset: GSE142564 (Day 0 and Day 14 bone marrow samples, mouse)
Language: R
Analysis Packages: Seurat, Monocle3, CellChat

---

Overview

This repository provides a complete, teaching-oriented single-cell RNA sequencing (scRNA-seq) tutorial. It integrates executable R analysis code with conceptual knowledge extracted from an accompanying lecture presentation, spanning 109 slides. The tutorial covers the full analytical trajectory — from raw 10X Genomics count matrices through quality control, normalization, data integration, dimensionality reduction, cell clustering, trajectory inference, and cell–cell communication analysis.

The biological context is a bone marrow immune profiling study comparing two time points (Day 0 vs. Day 14), with a particular focus on T-cell heterogeneity and intercellular signalling dynamics in what appears to be a B-cell precursor acute lymphoblastic leukaemia (BCP-ALL) or related haematopoietic model system.

---

Tutorial Goals

1. Understand the conceptual differences between bulk RNA-seq and scRNA-seq.
2. Learn how 10X Genomics droplet-based capture generates count matrices.
3. Apply rigorous, multi-metric quality control to exclude low-quality cells.
4. Perform SCTransform normalization with cell-cycle regression.
5. Integrate multi-sample data using Seurat's anchor-based integration framework.
6. Execute dimensionality reduction (PCA → UMAP) and assess clustering stability with clustree.
7. Annotate cell clusters using canonical lineage marker genes.
8. Subset and re-cluster a T-cell population; infer developmental trajectories with Monocle3.
9. Quantify and compare cell–cell communication networks across time points using CellChat.

---

Repository Structure

scrnaseq-tutorial/
├── README.md                          ← This file
├── LICENSE                            ← MIT License
├── .gitignore                         ← R/RStudio ignores
├── data/
│   └── README.md                      ← Data acquisition instructions (GEO: GSE142564)
├── code/
│   └── analysis_code.txt              ← Original analysis code (preserved)
├── docs/
│   ├── slide_by_slide_knowledge.md    ← Slide-by-slide extracted knowledge
│   ├── code_explanation.md            ← Section-by-section code walkthrough
│   ├── tutorial_workflow.md           ← Full sequential tutorial workflow
│   └── github_project_summary.md     ← Project summary for sharing
├── slides/
│   └── 10_slide_teaching_deck.md      ← Condensed 10-slide teaching deck
└── environment/
    └── requirements_or_dependencies.md ← R packages and dependencies

---

How to Use This Project

For Learners

1. Start with the concepts: Read docs/slide_by_slide_knowledge.md to build foundational understanding of scRNA-seq theory.
2. Follow the workflow: Work through docs/tutorial_workflow.md step by step, running code as you go.
3. Understand the code: Use docs/code_explanation.md alongside the actual code in code/analysis_code.txt.
4. Use the teaching deck: slides/10_slide_teaching_deck.md is a condensed lecture guide for presenting or self-teaching.

For Instructors

• The 10-slide deck in slides/ can be adapted into presentation software.
• The docs/tutorial_workflow.md provides a structured lab exercise template.
• Each code section is independently runnable if input data are available.

---

Expected Input Files

The analysis uses 10X Genomics CellRanger output files from GEO accession GSE142564:

GSE142564_RAW/
├── Day0/
│   ├── barcodes.tsv.gz
│   ├── features.tsv.gz
│   └── matrix.mtx.gz
└── Day14/
    ├── barcodes.tsv.gz
    ├── features.tsv.gz
    └── matrix.mtx.gz

Download from: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE142564

---

Analysis Workflow Summary

Stage	Key Action	Primary Tool
1. Data Loading	Read 10X CellRanger output	Seurat::Read10X()
2. QC (per sample)	Filter low-quality cells	Seurat::subset()
3. Metadata	Organise cell barcodes and metrics	tidyverse
4. Merge	Combine Day0 + Day14	Seurat::merge()
5. Gene Filtering	Remove Ig/TCR genes; sparse genes	Base R / Matrix
6. Cell Cycle	Score and regress S/G2M phases	Seurat::CellCycleScoring()
7. Normalisation	SCTransform per sample	Seurat::SCTransform()
8. Integration	Anchor-based cross-sample integration	Seurat::IntegrateData()
9. Dim. Reduction	PCA → UMAP	Seurat::RunPCA/RunUMAP()
10. Clustering	Leiden community detection + clustree	Seurat::FindClusters() + clustree
11. Annotation	Marker genes + FindAllMarkers	Seurat::FindAllMarkers()
12. Subclustering	T-cell subset analysis	Seurat
13. Trajectory	Pseudotime ordering	monocle3
14. Cell Comms	Ligand–receptor network comparison	CellChat

---

Who This Project Is For

• Graduate students and postdocs beginning scRNA-seq analysis
• Bioinformatics course instructors seeking a structured tutorial template
• Immunologists and cancer biologists learning computational methods
• R users transitioning from bulk RNA-seq to single-cell approaches

---

Learning Outcomes

After completing this tutorial, learners will be able to:

• Explain why scRNA-seq reveals biology that bulk RNA-seq cannot
• Apply and interpret multi-metric QC filtering decisions
• Understand the rationale for SCTransform over log-normalisation in multi-sample settings
• Implement Seurat anchor-based integration to remove batch effects
• Interpret UMAP embeddings and clustree stability plots
• Annotate clusters using canonical marker genes and FindAllMarkers()
• Perform subset re-clustering and trajectory inference on a cell population of interest
• Construct and compare CellChat communication networks across experimental conditions

---

Dataset: GEO accession GSE142564