GSE305595 Figure Reproduction

Code to reproduce Figures 1–6 for the manuscript: A Single-Cell Transcriptomic Atlas Identifies Hierarchical Skeletal Progenitors and Zonal Patterning in Mouse Mandibular Cartilage (Cell Reports).

Abstract

This repository contains the full data-processing and visualization pipeline used to generate the scRNA-seq figure panels for the manuscript. To maintain a lightweight repository size, large intermediate data objects (such as Seurat RDS files and scanpy H5AD files) are not included. Users must download the raw objects from GEO/Zenodo, place them in the correct directory structure, and execute the final baking script before running the plotting scripts.

---

1. Setup & Working Directory

IMPORTANT: All scripts in this repository use relative paths. Before running any scripts, you MUST set your working directory to the root of this repository.

In R:

setwd("/path/to/GSE305595")

---

2. Data Acquisition & Directory Structure

Download the processed datasets from the public repositories:

• GEO: GSE305595
• Zenodo: 10.5281/zenodo.20301128

Place the downloaded files into the data/raw/ directory EXACTLY as shown below:

GSE305595/
├── scripts/
├── output/                     # Generated figures will be saved here
└── data/
    ├── Final_Annotated_RDS/    # Will be populated by Step 1
    └── raw/
        ├── TMJ_INT_ALL.rds                 # <- Download from Zenodo
        ├── TMJ_13W_Mes.rds                 # <- Download from Zenodo
        ├── TMJ_P0_Mes.rds                  # <- Download from Zenodo
        ├── TMJ_E16_Mes.rds                 # <- Download from Zenodo
        ├── TMJ_E16_Chondro.rds             # <- Download from Zenodo
        ├── TMJ_E16_Chondro_scVelo/
        │   └── scVelo.h5ad                 # <- Download from Zenodo
        ├── TMJ_E16_Chondro_STREAM/
        │   └── STREAM_out.pkl              # <- Download from Zenodo
        └── [Mapping CSVs included in repo]

---

3. Dependencies

R Environment

• R ≥ 4.3.0
• Seurat
• SeuratWrappers
• SeuratDisk
• dplyr, ggplot2, RColorBrewer, scales, viridis, ggrepel, gridExtra, stringr, openxlsx
• ComplexHeatmap, circlize
• clusterProfiler, org.Mm.eg.db
• CellChat

Python Environments

• scVelo/Scanpy: scvelo, scanpy, pandas, numpy, matplotlib
• STREAM: We recommend a dedicated conda environment with stream installed.

---

4. Quick Start Execution Order

Run the scripts in the following order.

Step 1: Bake Annotated Datasets

source("scripts/99_bake_final_rds.R")

(This applies biological cluster mappings and factor levels, saving the final objects to data/Final_Annotated_RDS/)

Step 2: Generate Figures

source("scripts/01_fig1_atlas.R")       # Fig 1A-C
source("scripts/01_fig1_FEA.R")         # Fig 1D-F
source("scripts/02_fig2_13w.R")         # Fig 2
source("scripts/03_fig3_P0.R")          # Fig 3
source("scripts/04_fig4_E16.R")         # Fig 4
source("scripts/05_fig5_trajectory.R")  # Fig 5A-F (Calls Python scripts for 5C, 5G, 5H)
source("scripts/06_fig6e_crabp1.R")     # Fig 6E

(Optional: Run 07_cellchat.R to regenerate CellChat analysis. This is computationally intensive.)

---

5. Figure Index

Figure	Script	Output File
Fig 1A	01_fig1_atlas.R	Fig1/Fig1A_UMAP_Stage.png
Fig 1B	01_fig1_atlas.R	Fig1/Fig1B_UMAP_Cluster.png
Fig 1C	01_fig1_atlas.R	Fig1/Fig1C_FeaturePlot_*.png
Fig 1D	01_fig1_FEA.R	Fig1/Fig1D_E16.5_GO_DotPlot.png
Fig 1E	01_fig1_FEA.R	Fig1/Fig1E_P0_GO_DotPlot.png
Fig 1F	01_fig1_FEA.R	Fig1/Fig1F_13w_GO_DotPlot.png
Fig 2	02_fig2_13w.R	Fig2/*.png
Fig 3	03_fig3_P0.R	Fig3/*.png
Fig 4	04_fig4_E16.R	Fig4/*.png
Fig 5A	05_fig5_trajectory.R	Fig5/Fig5A_UMAP.png
Fig 5B	05_fig5_trajectory.R	Fig5/Fig5B_FeaturePlot_*.png
Fig 5C	05_fig5_scvelo.py	Fig5/Fig5C_scVelo_Stream.png
Fig 5D	05_fig5_trajectory.R	Fig5/Fig5D_UMAP_Pseudotime.png
Fig 5E	05_fig5_trajectory.R	Fig5/Fig5E_FeaturePlot_*.png
Fig 5F	05_fig5_trajectory.R	Fig5/Fig5F_DotPlot.png
Fig 5G	05_fig5_stream.py	Fig5/Fig5G_STREAM_Branches.png
Fig 5H	05_fig5_stream.py	Fig5/Fig5H_STREAM_SubwayMap.png, Fig5H_STREAM_*.png
Fig 6E	06_fig6e_crabp1.R	Fig6/Fig6E_Crabp1_DotPlot.png

---

License

This code is released under the CC-BY-4.0 License.