SlimR: Marker-Based R Package for Single-Cell and Spatial-Transcriptomic Annotation

Overview

SlimR is an R package designed for annotating single-cell and spatial-transcriptomics (ST) datasets. It supports the creation of a unified marker list (‘Markers_list’) using multiple sources including: user-provided Excel tables mapping cell types to markers, Seurat objects containing cell label information, and the package’s built-in curated species-specific cell type and marker reference databases (e.g., ‘Cellmarker2’, ‘PanglaoDB’).

Based on the Markers_list, ‘SlimR’ enables one-click generation of annotation heatmaps (‘Annotation_heatmap’) visualizing relationships between input cell types and reference marker lists. Additionally, it can iterate through different cell types to generate corresponding annotation reference plots (e.g., ‘Markers_dotplot’, ‘Metric_heatmap’, ‘Mean_expression_box_plot’).

Table of Contents

1. Preparation
   • 1.1 Installation
     
   • 1.2 Loading SlimR
     
   • 1.3 Dependencies
2. Standardized Marker_list Input
   • 2.1 From Excel Tables
     
   • 2.2 From Seurat Objects
     
   • 2.3 From Preprocessed Cellmarker2 Database
     
   • 2.4 From Preprocessed PanglaoDB Database
3. Automated Annotation Workflow
   • 3.1 Annotation Heatmap
     
   • 3.2 Annotation Box Plot
4. Semi-Automated Annotation Workflow
   • 4.1 With Excel-Based Marker Lists
     
   • 4.2 With Seurat-Based Marker Lists
     
   • 4.3 With Cellmarker2 Database
     
   • 4.4 With PanglaoDB Database
5. Conclusion

1. Preparation

1.1 Installation

Install SlimR directly from GitHub using:

devtools::install_github("Zhaoqing-wang/SlimR")

1.2 Loading SlimR

Load the package in your R environment:

library(SlimR)

1.3 Dependencies

SlimR requires R (≥ 3.5) and depends on the following packages: cowplot, dplyr, ggplot2, patchwork, pheatmap, readxl, scales, Seurat, tidyr, tools. Install missing dependencies using:

# Install dependencies if needed:
install.packages(c("cowplot", "dplyr", "ggplot2", "patchwork", 
                   "pheatmap", "readxl", "scales", "Seurat", 
                   "tidyr", "tools"))

2. Standardized Marker_list Input

SlimR requires a standardized list format for storing marker information, metrics, and corresponding cell types (list names = cell types, first column = markers, subsequent columns = metrics).

2.1 From Excel Tables

Format Requirements:
- Each sheet name = cell type
- First row = column headers
- First column = markers
- Subsequent columns = metrics

Markers_list_Excel <- read_excel_markers("D:/Laboratory/Marker_load.xlsx")

Note: Output usable in sections 3.1, 3.2, and 4.1.

2.2 From Seurat Objects

First identify cluster features:

seurat_markers <- FindAllMarkers(
  sce.all, 
  group.by = "Cell_type", 
  only.pos = TRUE
)

Then generate marker list:

Markers_list_Seurat <- read_seurat_markers(
  seurat_markers,
  sort_by = "avg_log2FC",
  gene_filter = 10
)

Note: Output usable in sections 3.1, 3.2, and 4.2.

2.3 From Preprocessed Cellmarker2 Database

Load the database:

Cellmarker2 <- SlimR::Cellmarker2

Optional metadata exploration:

Cellmarker2_table <- SlimR::Cellmarker2_table
View(Cellmarker2_table)

Generate marker list:

Markers_list_Cellmarker2 <- Markers_filter_Cellmarker2(
  Cellmarker2,
  species = "Human",
  tissue_class = "Intestine",
  tissue_type = NULL,
  cancer_type = NULL,
  cell_type = NULL,
  cell_name = NULL,
  marker = NULL,
  counts = NULL
)

Note: Output usable in sections 3.1, 3.2, and 4.3.

2.4 From Preprocessed PanglaoDB Database

Load the database:

PanglaoDB <- SlimR::PanglaoDB

Optional metadata exploration:

PanglaoDB_table <- SlimR::PanglaoDB_table
View(PanglaoDB_table)

Generate marker list:

Markers_list_panglaoDB <- Markers_filter_PanglaoDB(
  PanglaoDB,
  species_input = 'Human',
  organ_input = 'GI tract'
)

Note: Output usable in sections 3.1, 3.2, and 4.4.

3. Automated Annotation Workflow

3.1 Annotation Heatmap

Generates a heatmap comparing marker expression across cell clusters:

Celltype_annotation_Heatmap(
  seurat_obj = sce,
  gene_list = Markers_list,
  species = "Human",
  cluster_col = "seurat_cluster",
  min_expression = 0.1,
  specificity_weight = 3
)

3.2 Annotation Box Plot

Generates per-cell-type expression box plots:

Celltype_annotation_Box(
  seurat_obj = sce,
  gene_list = Markers_list, 
  species = "Human",
  cluster_col = "seurat_cluster",
  assay = "RNA",
  save_path = "./SlimR/Celltype_annotation_Bar/"
)

4. Semi-Automated Annotation Workflow

4.1 With Excel-Based Marker Lists

Generates integrated dot plots and metric heatmaps:

Celltype_annotation_Excel(
  seurat_obj = sce,
  gene_list = Markers_list_Excel,
  species = "Human",
  cluster_col = "seurat_cluster",
  assay = "RNA",
  save_path = "./SlimR/Celltype_annotation_Excel/"
)

4.2 With Seurat-Based Marker Lists

Celltype_annotation_Seurat(
  seurat_obj = sce,
  gene_list = Markers_list_Seurat,
  species = "Human",
  cluster_col = "seurat_cluster",
  assay = "RNA",
  save_path = "./SlimR/Celltype_annotation_Seurat/"
)

4.3 With Cellmarker2 Database

Celltype_annotation_Cellmarker2(
  seurat_obj = sce,
  gene_list = Markers_list_Cellmarker2,
  species = "Human",
  cluster_col = "seurat_cluster",
  assay = "RNA",
  save_path = "./SlimR/Celltype_annotation_Cellmarkers2.0/"
)

4.4 With PanglaoDB Database

Celltype_annotation_PanglaoDB(
  seurat_obj = sce,
  gene_list = Markers_list_panglaoDB,
  species = "Human",
  cluster_col = "seurat_cluster",
  assay = "RNA",
  save_path = "./SlimR/Celltype_annotation_PanglaoDB/"
)

5. Conclusion

Thank you for using SlimR. For questions, issues, or suggestions, please contact:

Zhaoqing Wang
📧 851091628@qq.com; zhaoqingwang@mail.sdu.edu.cn