Enteric Neuron Subtype Analysis Workflow

This repository provides a workflow for analyzing enteric neuron subtypes using single-cell RNA sequencing (scRNA-seq) data processed with the Seurat framework. It accompanies the manuscript Discovering the transcriptomes, connections, and development of submucosal neuron classes in the mouse small intestine (Li and Morarach et al., 2024). The associated data is deposited in the Gene Expression Omnibus (GEO) database under the identifier GSE263422.

The analysis procedures described in the manuscript are wrapped into a series of custom functions with adjustable parameters for reproducibility and ease of use. The main script, RSCRIPT.R, is organized into three parts:

• Part 1: Libraries – Load required R packages.
• Part 2: Functions – Define custom functions for preprocessing and analysis.
• Part 3: Analyses – Perform the analyses, organized into four main sections.
• (1) General Workflow for Cluster Analysis (Compatible with Seurat v5)
• (2) P24 Submucosal Plexus Cluster Analysis
• (3) P7 Small Intestine (ENS) Cluster Analysis
• (4) Label Transfer

To follow this workflow, first load the Libraries and Functions sections of the RSCRIPT.R file. These only need to be loaded once per session. The Analyses section can then be executed as needed.

Download data:

• Raw Data for P24 Submucosal Plexus:
  GSM8191472,
  GSM8191473,
  GSM8191474
• Raw Data for P7 Small Intestine ENS:
  GSM8191470,
  GSM8191471
• Processed P24 submucosal plexus
• Processed P7 small intestine ENS

---

Analysis Overview

Main functions

For details, refer to the comments and descriptions embedded in the script.

1. Data Loading
   • load_10x(): Load single-cell RNA sequencing data from 10x Genomics.
2. Pre-processing
   • filter_cell_by_threshold(): Exclude low-quality cells based on thresholds for nFeature_RNA, nCount_RNA and percent.mt.
   • get_cluster_outlier(): Exclude extreme outliers, sweeping different resolution values.
   • run_dbfinder(): Exclude doublets by DoubletFinder (with Seurat’s SCT)
3. Iterative Clustering and removing of non-relevant clusters
   • analyze_sctseurat(): Identify and subset non-enteric neuron clusters iteratively.

---

Example of Dataset loading

Specify the folders containing paths to the CellRanger's output, filtered_feature_bc_matrix:

folders <- list(
  "~/sm022/filtered_feature_bc_matrix",
  "~/sm023/filtered_feature_bc_matrix",
  "~/sm024/filtered_feature_bc_matrix"
)
load_sm <- load_data_10x(folders, seurat_object_version = "v5") #for Seurat v5

---

Library Requirements

• R 4.4.1
• Python 3.13.0

This script works with the following package versions:

• cols4all: v0.7-1
• reticulate: v1.39.0
• patchwork: v1.3.0
• clustree: v0.5.1
• ggraph: v2.2.1
• ggplot2: v3.5.1
• DoubletFinder: v2.0.4
• dplyr: v1.1.4
• Seurat: v5.1.0
• SeuratObject: v5.0.2
• sp: v2.1-4
• stringr: v1.5.1

Seurat (via reticulate) also uses the following Python packages:

• umap-learn: v0.5.7
• leidenalg: v0.10.2

Reproducibility Requirements

To reproduce the results, specific package versions are required:

• Seurat: v4.1.3
• DoubletFinder: v2.0.3 (for the function paramSweep_v3() and doubletFinder_v3())

---

Citation

Discovering the transcriptomes, connections, and development of submucosal neuron classes in the mouse small intestine.
Wei Li,#, Khomgrit Morarach,#, Ziwei Liu, Sanghita Banerjee, Yanan Chen,  Ashley L. Harb, Joel M. Kosareff, Charles R. Hall, Fernando López-Redondo, Elham Jalalvand, Suad H. Mohamed, Anastassia Mikhailova, David R. Linden, Ulrika Marklund,*

#These authors contributed equally
*Correspondence:
Ulrika Marklund (PhD), Karolinska Institute Department of Medical Biochemistry and Biophysics, Division of Molecular Neurobiology, Biomedicum C6, Tomtebodavägen 16, SE-171 65, Stockholm, Sweden. Email: Ulrika.Marklund@ki.se
Telephone: +46 702931870

---