🧬 CyteTypeR Tutorial

A hands-on learning guide for CyteTypeR — an R package that automatically annotates your scRNA-seq clusters using the CyteType API.

TL;DR — Feed it a Seurat object with cluster markers, get back cell type annotations + a beautiful interactive report. ✨

🔬 What It Does

Seurat Object ──→ PrepareCyteTypeR() ──→ CyteTypeR API ──→ Annotated Object + HTML Report
     │                    │                                        │
  clusters            extracts                              predicted cell
  markers            markers,                               types added to
  UMAP              metadata,                                 metadata
                   coordinates

📦 Requirements

# Core
install.packages(c("Seurat", "tidyverse"))

# CyteTypeR
devtools::install_github("NygenAnalytics/CyteTypeR")

# Dependencies
install.packages(c("SCpubr", "glmGamPoi", "tictoc"))
remotes::install_github("satijalab/seurat-data")

🚀 Quick Start

library(Seurat)
library(CyteTypeR)

# 1. Load and preprocess
obj <- LoadData("ifnb")
obj <- SCTransform(obj, vst.flavor = "v2")
obj <- RunPCA(obj)
obj <- RunUMAP(obj, dims = 1:20)
obj <- FindNeighbors(obj, dims = 1:20)
obj <- FindClusters(obj, resolution = 0.4)

# 2. Find markers
markers <- FindAllMarkers(obj, group.by = "seurat_clusters")

# 3. Prepare for CyteTypeR
prepped_data <- PrepareCyteTypeR(obj, markers, n_top_genes = 10)

# 4. Annotate 🎯
annotated_obj <- CyteTypeR(obj, prepped_data, study_context = "PBMC IFNB stimulation")

🚀 How to Use

1. Clone this repo
2. Open cytetype.Rproj in RStudio (this sets your working directory and ensures reproducibility)
3. Open and run through cytetyper_guide.qmd step by step

📂 Files

File	Description
cytetype.Rproj	🔧 RStudio project file — open this first
cytetyper_guide.qmd	📖 Full step-by-step tutorial (Quarto)
query.json	🔍 Example API request/response structure

🔗 Resources

	Link
🧪 CyteTypeR (R)	github.com/NygenAnalytics/CyteTypeR
🐍 CyteType (Python)	github.com/NygenAnalytics/CyteType
📊 Example Report	Interactive HTML Report

---

Made with ☕ while learning single-cell bioinformatics.