Comprehensive Single-Cell Transcriptomic Atlas of Mouse Pons & Medulla

This repository provides metadata, scripts, and documentation for the large-scale transcriptomic atlas of the mouse pons and medulla subregions, compiled from 8 independent single-cell/single-nucleus RNA sequencing (sc/snRNA-seq) datasets.

The integration was performed using a standardized bioinformatic workflow, clustering, and marker-based annotation. The dataset comprises 317,985 quality-passed cells across 45 cell types.

This integrated large-scale dataset serves as a valuable single-cell neuroscience dataset to explore region-specific molecular insights and cellular diversity in the brainstem.

---

Dataset Overview

• Cells: 317,985 quality passed cells
• Species: Mouse (Mus musculus)
• Cell Types: 45 distinct cell types
• Methodology: scRNA-seq/snRNA-seq integration
• Data Format:.rds files (Seurat objects)
• Integration Tool: Seurat v4.4.2

Download the Dataset

The .rds files containing normalized expression data can be accessed via Figshare:

1. Final_Integrated_Pons_Medulla.rds | Full dataset with all identified cell types |Download](https://doi.org/10.6084/m9.figshare.28342025.v4) |
2. Pons_Medulla_Neurons_level3.rds | Neuronal subtypes dataset | Download
3. Differentially Expressed Genes (DEGs)

File: DEG_Level1_Pons_Medulla.csv

Description: Cluster-level transcriptional differences across all cells (Level 1 annotations).

File: DEG_Level1_Neurons.csv

Description: Cluster-level transcriptional differences within neuronal populations only.

Reproducing the Analysis

Code and Technical Details

All technical details for:

Preprocessing

Integration

Clustering

Cell type annotation

Subclustering and subtyping

are provided in the code section of this repository.

Prerequisites

• R version: ≥ 4.2.1
• Key R packages:
  • devtools
  • reticulate
  • Seurat
  • Matrix
  • dplyr
  • ggplot2
  • tidyr
  • tidyverse
  • readxl
  • ComplexHeatmap
  • RColorBrewer
  • clusterProfiler

Install dependencies in R:

install.packages(c("devtools", "reticulate", "Seurat", "Matrix", "dplyr", 
                   "ggplot2", "tidyr", "tidyverse", "readxl", "ComplexHeatmap", 
                   "RColorBrewer"))
if (!requireNamespace("BiocManager", quietly = TRUE))
    install.packages("BiocManager")
BiocManager::install("clusterProfiler")

## How to Use the Data
1. Download the `.rds` files from Figshare.
2. Load the dataset into R:
   ```r
   library(Seurat)
   pons_medulla <- readRDS("Final_Integrated_Pons_Medulla.rds")

# Check metadata
head(pons_medulla@meta.data)

# View the annotated UMAP
DimPlot(pons_medulla, reduction = "umap", group.by = "cell_type")

# View all annotated cell types
unique(pons_medulla$cell_type)

# Cell counts per dataset
table(pons_medulla$Study)

# Cell counts per cell type
table(pons_medulla$cell_type)

# Final UMAP with all major cell types and subtype

DimPlot(pons_medulla, reduction = "umap", group.by = "cell_type", raster = FALSE, pt.size = 0.2, label = FALSE) + 
  scale_color_manual(values = custom_colors) 

# Cell type proportion group by each study
# data frame for cell type counts
cell_type_counts <- as.data.frame(table(pons_medulla$cell_type))
colnames(cell_type_counts) <- c("Cell_Type", "Count")

# Bar plot
ggplot(cell_type_counts, aes(x = reorder(Cell_Type, -Count), y = Count, fill = Cell_Type)) +
  geom_bar(stat = "identity") +
  scale_fill_manual(values = custom_colors) +  # Use your defined colors
  theme_minimal() +
  theme(axis.text.x = element_text(angle = 45, hjust = 1, size = 10),  # Adjust text
        axis.title = element_text(size = 12, face = "bold"),
        legend.position = "none") +
  labs(title = "Cell Type Distribution", x = "Cell Type", y = "Count")