/media/jhuang/Elements(Denise_ChIPseq)/Data_Jingang/GSE163973_KF_NS_done/seurat0**.Rmd[R]
https://nbisweden.github.io/workshop-scRNAseq/
https://github.com/NBISweden/workshop-scRNAseq/
https://github.com/NBISweden/workshop-scRNAseq/blob/master/labs/seurat/seurat_06_celltyping.qmd
https://github.com/NBISweden/workshop-scRNAseq/blob/master/labs/seurat/seurat_07_trajectory.qmd
Project Workflow
The analysis is organized into seven steps, each corresponding to an R Markdown file:
- 01_qc – Quality control of the raw data.
- 02_dimension_reduction – Dimensionality reduction (e.g., PCA, UMAP).
- 03_integration – Data integration using the Harmony library to align datasets and mitigate batch effects. This step also includes normalization (
NormalizeData) and selection of highly variable genes (FindVariableFeatures) for each dataset before integration. - 04_clustering – Identification of cell clusters.
- 05_dge – Differential gene expression analysis.
- 06_celltype – Cell type annotation.
- 07_trajectory
- 08_spatial – if applicable
This structure should help you follow the workflow step by step.
Harmony in R: Integration vs. Batch Effect Removal
Harmony is a tool commonly used in single-cell RNA-seq analysis. Its main purpose is data integration, but it also effectively removes batch effects. Here’s a breakdown:
1. What Harmony Does
- Aligns multiple datasets (from different batches, labs, conditions, or technologies) in a shared low-dimensional space (e.g., PCA).
- Reduces technical variation (batch effects) while preserving biological differences.
- Produces a “corrected” dataset suitable for downstream analysis like clustering or visualization.
2. Integration vs. Batch Effect Removal
| Term | Meaning |
|---|---|
| Batch Effect Removal | Focuses only on removing technical variation between batches. May distort biological differences. |
| Integration | Aligns datasets from different batches or conditions, minimizing batch effects while keeping biological variation intact. Harmony achieves this by iteratively adjusting cells’ embeddings. |
3. Conceptual Diagram
Batch 1: A1 A2 A3 →
Batch 2: B1 B2 B3 —-> [Harmony] —> Integrated Space (similar cell types cluster together) Batch 3: C1 C2 C3 →
- Arrows indicate the mapping of cells from separate batches into a shared low-dimensional space.
- Cells of the same type cluster together, regardless of batch.
4. Summary
- Harmony = integration tool.
- Batch effect removal = part of the integration process.
- Integration = alignment of datasets with biological signals preserved.