Mouse Skin Single Cell Rna Seq

Single cell RNA sequencing (scRNA seq) of mouse skin is a method that profiles gene expression in thousands of individual cells isolated from the epidermis, dermis, and subcutaneous layers, enabling discovery of rare cell types, developmental trajectories, and disease associated transcriptional programs. This guide is intended for graduate students, postdoctoral researchers, and principal investigators who plan to design or interpret a scRNA seq experiment focused on mouse skin. It provides a source bounded framework to navigate core concepts, make informed decisions, avoid common errors, and understand the inherent limits of the technique. For a broad introduction to scRNA seq principles, see the EMBL EBI Training materials. The authoritative technical background on sequencing methodologies can be found in the NCBI Bookshelf reference collection.

At a Glance

Core Concepts and Decision Points

Why Mouse Skin

Mouse skin is a widely used model for studying development, wound healing, inflammation, and carcinogenesis because of its genetic tractability and similarity to human skin architecture. The tissue contains a stratified epidermis, hair follicles, sebaceous glands, and a complex dermis with fibroblasts, blood vessels, and resident immune populations. A single cell transcriptome can reveal how these compartments communicate and respond to perturbations. However, the skin is resilient and requires robust dissociation protocols. For guidance on computational analysis pipelines, consult the Galaxy Training Network workflow tutorials.

Decision Points Before Starting

1. Dissociation method. Over digestion damages cells and increases stress related transcripts. Under digestion leaves cell clumps that cause doublet artifacts. A common protocol uses a mix of collagenase I and dispase for adult mouse skin, followed by filtration.
2. Library type. Droplet based methods capture 3 prime ends and allow thousands of cells per sample at lower cost per cell. Plate based methods sequence full length transcripts with higher sensitivity, which can help detect lowly expressed genes but are limited to hundreds of cells.
3. Number of replicates. At least three biological replicates per condition are recommended to capture biological variability. Pooling mice may be necessary if individual skin samples yield few cells.
4. Sequencing depth. For cell type identification, 20,000 to 50,000 reads per cell are typical. Deeper sequencing (100,000+ reads) aids splice variant or allelic expression detection.

These decisions affect downstream resolution. Detailed analysis software and workflows are provided by Bioconductor with packages such as scran, iSEE, and SingleCellExperiment.

Practical Workflow

The workflow below assumes a droplet based 10x Genomics or similar platform, but many steps apply generally.

1. Tissue collection and dissociation. Harvest mouse skin (usually dorsal or ear), remove subcutaneous fat, mince into 1 2 mm pieces, digest in enzyme cocktail for 30 90 minutes at 37 degrees Celsius with gentle agitation. Quench with serum containing medium, filter through 70 micron and 40 micron strainers, wash, and count viable cells. Aim for > 80% viability.
2. Library preparation and sequencing. Follow the manufacturer protocol for single cell 3 prime or 5 prime gene expression. Sequence on a NovaSeq or NextSeq to achieve the desired read depth. Raw sequencing data in BCL format is demultiplexed to FASTQ files.
3. Preprocessing. Use Cell Ranger (10x) to align reads to the mm10 genome, collapse UMIs, and generate gene cell count matrices. Alternative tools like STARsolo can be used. Raw data must be deposited in public repositories such as the NCBI Sequence Read Archive.
4. Quality control and filtering. Remove cells with fewer than 200 genes or more than 7000 genes, and those with more than 10% mitochondrial reads. These thresholds filter out empty droplets, doublets, and dying cells. Doublet scoring algorithms (e.g., DoubletFinder) help identify heterotypic doublets.
5. Normalization and scaling. Normalize raw counts with library size factors (log transformation) or using the scran deconvolution method. Scale data to unit variance per gene.
6. Dimensionality reduction and clustering. Perform PCA on highly variable genes (top 2000). Use Harmony or scVI to integrate samples if batch effects are present. Then run UMAP for visualization and graph based clustering (Leiden or Louvain).
7. Cell type annotation. Identify clusters by expression of known markers: keratin 14 and keratin 5 for basal keratinocytes, keratin 10 for spinous keratinocytes, Col1a1 and Dcn for fibroblasts, Cd45 or Ptprc for immune cells, etc. Reference based mapping tools (SingleR) can aid annotation using datasets like the Tabula Muris.
8. Differential expression and downstream analysis. Compare clusters across conditions using pseudobulk methods (e.g., DESeq2 on aggregated counts) or mixed models (e.g., MAST). Trajectory inference (Monocle 3, Slingshot) can map differentiation pathways such as keratinocyte maturation. Check for cell cell communication using CellChat or NicheNet.

For a real world example, a study on atopic dermatitis identified a MIF driven keratinocyte macrophage Th2 axis using scRNA seq of mouse skin. See Single cell transcriptomic analysis identifies a MIF driven keratinocyte macrophage Th2 axis.

Quality Checks

High quality scRNA seq data meets these benchmarks:

• Median genes per cell: 1000 3000 for 10x data.
• Median UMI count per cell: 2000 10,000.
• Fraction of mitochondrial reads: < 10% (ideally < 5%).
• Doublet rate: < 5% after filtering.
• Number of cells per sample: > 1000 viable cells after QC.
• Low ambient RNA contamination (e.g., from lysed cells). SoupX or DropletUtils can estimate and subtract ambient RNA.

Always visualize these metrics per sample before merging. A quality control report with plots (nGene vs nUMI, mito% vs nGene) should be generated.

Common Mistakes

1. Using the same dissociation protocol for all ages or conditions. Neonatal skin is more fragile than adult skin, over digestion destroys cells and enriches for stress resilient subpopulations. Titrate enzyme for each experiment.
2. Ignoring batch effects. Processing samples on different days or with different reagent lots introduces technical variation. Always include replicates and use integration algorithms.
3. Annotating cells solely by high expression of a single marker. For example, Cd45 marks all immune cells, but not all immune cells express it at detectable levels. Use multiple markers per cell type.
4. Failing to check for doublets. Heterotypic doublets can appear as false intermediate cell types. Always run doublet detection and inspect clusters co expressing markers of two distinct lineages.
5. Overinterpreting rare clusters. A cluster with fewer than 20 cells may be artifact or a small subpopulation, validate with independent methods like flow cytometry or in situ hybridization.

Limits and Uncertainty

ScRNA seq data is sparse (dropout events) because of low capture efficiency. Many genes expressed at low levels are not detected. This limits analysis of weak transcriptional changes. Additionally, tissue dissociation induces a transcriptional stress response that can mask subtle biological differences. Spatial context is lost, neighboring cell interactions must be inferred computationally. For spatial resolution, consider coupling with spatial transcriptomics as demonstrated in a study profiling ten mouse organs A unified spatial transcriptome profiling of ten mouse organs.

Single cell data also requires careful statistical interpretation. Differential expression tests may produce false positives due to zero inflation. Pseudobulk aggregation helps but reduces cell level resolution. Finally, while mouse skin is a powerful model, findings may not directly translate to human skin due to differences in immune composition and barrier function. Recent work on EMT tumor states in skin highlights these complexities Single cell multiomics unravel the transcription networks controlling the different EMT tumor states.

Frequently Asked Questions

What is the best dissociation method for mouse skin?

There is no universal best method. For adult dorsal skin, a 60 minute digestion with 1 mg/ml collagenase I and 2.5 mg/ml dispase at 37 degrees Celsius works well. Always perform a viability check and run a test experiment to optimize.

How many cells should I aim to sequence per sample?

For cell type discovery, 5,000 10,000 cells per sample are often sufficient. For rare populations (e.g., resident memory T cells), 15,000 20,000 cells are recommended.

How can I identify immune cells in mouse skin scRNA seq data?

Look for expression of Ptprc (Cd45), then subset and recluster. Use markers like Cd3e (T cells), Nkg7 (NK cells), Cd68/Mrc1 (macrophages), and Fcgr1 (monocytes). Correlate with literature such as the study on psoriatic inflammation PRMT5 promotes psoriatic inflammation through mediating dimethylation of p65 in keratinocytes.

What are common artifacts in skin scRNA seq?

Doublets between keratinocytes and fibroblasts are frequent. Ambient RNA from dead cells can make clusters appear homogeneous. Also, stress response genes (JUN, FOS, HSP genes) can dominate variance if dissociation is harsh.

References and Further Reading

1. NCBI Bookshelf. Comprehensive technical background on molecular biology and sequencing. https://www.ncbi.nlm.nih.gov/books/
2. EMBL EBI Training. Tutorials on single cell RNA seq analysis. https://www.ebi.ac.uk/training/
3. Galaxy Training Network. Workflow based bioinformatics training for scRNA seq. https://training.galaxyproject.org/
4. Bioconductor. Software and guides for single cell genomic analysis. https://bioconductor.org/
5. NCBI Sequence Read Archive. Repository for raw sequencing data. https://www.ncbi.nlm.nih.gov/sra
6. Single cell multiomics unravel the transcription networks controlling the different EMT tumor states. Nat Commun. https://pubmed.ncbi.nlm.nih.gov/42436193/
7. PRMT5 promotes psoriatic inflammation through mediating dimethylation of p65 in keratinocytes. Mol Med. https://pubmed.ncbi.nlm.nih.gov/42432494/
8. Single cell transcriptomic analysis identifies a MIF driven keratinocyte macrophage Th2 axis as a key inflammatory circuit in atopic dermatitis. Inflamm Res. https://pubmed.ncbi.nlm.nih.gov/42406119/
9. A unified spatial transcriptome profiling of ten mouse organs. Sci Data. https://pubmed.ncbi.nlm.nih.gov/42374103/
10. Loss of epidermal miR 149 sensitizes to skin inflammation. J Allergy Clin Immunol. https://pubmed.ncbi.nlm.nih.gov/42362092/
11. Targeting LRRC15 expressing fibroblasts holds therapeutic potential to ameliorate inflammation and fibrosis in autoimmune diseases. Rheumatology (Oxford). https://pubmed.ncbi.nlm.nih.gov/42340684/

Related Articles

• Protein Synthesis
• Incomplete Dominance Gene
• Cell Membrane Function Biology
• Dna Structure
• Protein Structure