1  Quick Recap: Introduction to scRNA‑seq + CITE-seq

1.1 scRNA-seq vs bulk-RNA-seq

  • Bulk RNA‑seq: measures average gene expression in a tissue; useful for detecting global changes, but it hides cellular heterogeneity.

    • Insufficient for: early developmental studies or complex tissues such as the brain.

Image source: Yan et al. 2024. Int J Biol Sci

1.2 Single cell RNA-seq (scRNA-seq)

  • First publication: 2009 (Tang et al, 2009)
  • Popularity: 2014
  • Estimate the distribution of gene expression levels across a cell population.
  • Analysis at the single‑cell level.
  • Reveals cell types, including rare populations.
  • Analyze temporal dynamics and heterogeneity.

Image source: Single cell RNA sequencing

1.3 Heterogeneity: Internal diversity within a system or a set

  • 🔬 Not all cells are the same: even within the same tissue, there are differences in cell types, functional states, and gene expression levels.

  • 🧬 Genetic and transcriptomic variability: each cell may express different genes or varying amounts of RNA, reflecting specific functions.

  • 🩺 Relevance in disease: cellular heterogeneity explains why a tumor, for example, contains subpopulations with distinct behaviors (some more aggressive, others more sensitive to treatment).

Visualization through UMAP (Uniform Manifold Approximation and Projection):

Image source: Explicación completa de un UMAP

1.4 Questions scRNA‑seq can answer

  • Discover new or rare cell types.
  • Identify differential cellular composition between healthy vs. diseased samples.
  • Understand cell differentiation in development and regeneration.
  • Analyze plasticity and expression dynamics at the single‑cell level.
  • Build cellular/genetic atlases → comprehensive catalog of cellular diversity.
  • Applications in basic research and personalized medicine.

Image source: Kageyama, et al. 2018. Front. Neurosci

1.5 General Pipeline for scRNA‑seq with Seurat

Step Purpose Seurat Command
  1. Preprocessing
 Load raw count matrix and create Seurat object Read10X(), CreateSeuratObject()
  1. Quality Control (QC)
 Filter cells/genes based on metrics (e.g., nFeature_RNA, percent.mt) subset(), PercentageFeatureSet()
  1. Normalization
 Adjust counts for sequencing depth NormalizeData() or SCTransform()
  1. Feature Selection
 Identify variable genes FindVariableFeatures()
  1. Dimensionality Reduction
 Reduce dimensions for visualization ScaleData(), RunPCA(), RunUMAP(), RunTSNE()
  1. Clustering & Annotation
 Group cells and assign biological meaning FindNeighbors(), FindClusters(), FindAllMarkers()
  1. Dataset Integration
 Combine multiple datasets, remove batch effects FindIntegrationAnchors(), IntegrateData()
  1. Downstream Analysis
 Specific goals: DEGs, trajectories, cell communication FindMarkers(), CellCycleScoring(), external packages

1.6 Variations of scRNA‑seq

Type of Analysis Contribution
Unimodal (RNA only) Cellular heterogeneity in gene expression
Multimodal (RNA + others, e.g., CITE‑seq) Regulation, proteomics, and biological context
Trajectories Temporal dynamics of biological processes
Cellular Interactions Communication and functional networks
Dataset Integration Comparison across conditions

1.7 What is CITE-Seq?

CITE‑seq (Cellular Indexing of Transcriptomes and Epitopes by Sequencing) is a multimodal single‑cell technique that combines RNA sequencing with protein profiling at the level of individual cells.

1.7.1 📌 Key points about CITE‑seq

  • Simultaneous measurements → captures both the transcriptome (scRNA‑seq) and surface proteins using antibody‑derived tags (ADTs).
  • Highly multiplexed → allows detection of many protein markers in parallel, alongside unbiased RNA profiling.
  • Single‑cell resolution → thousands of cells can be analyzed at once, integrating RNA and protein data.
  • Multimodal context → CITE‑seq is part of the broader multimodal analysis frontier in single‑cell genomics, where multiple data types (RNA, ATAC, methylation, perturbations) are measured in the same cell.
  • Applications → improves cell type identification, refines clustering, and reveals functional states that RNA alone might miss.

Schematic figure of (A) the CITE-seq antibody linked with the barcoded oligo (B) the CITE-seq protocol. Credit: Winston & Gregory Timp (2020), license: CC-BY 4.0

1.8 Overview

Schematic of the dry lab workflow for CITE-Seq

1.9 General Pipeline for CITE‑seq

Step Purpose Seurat Command
  1. Preprocessing
 Load RNA counts and antibody‑derived tags (ADT) Read10X(), CreateSeuratObject()
  1. Quality Control (QC)
 Filter cells/genes based on RNA metrics and ADT counts PercentageFeatureSet(), subset()
  1. Normalization (RNA)
 Adjust RNA counts for sequencing depth SCTransform() or NormalizeData()
  1. Normalization (ADT)
 Center and scale protein counts NormalizeData(assay = "ADT"), ScaleData(assay = "ADT")
  1. Feature Selection (RNA)
 Identify variable genes FindVariableFeatures()
  1. Dimensionality Reduction
 Reduce dimensions for RNA and ADT separately RunPCA(assay = "RNA"), RunPCA(assay = "ADT")
  1. Multimodal Integration
 Combine RNA + ADT signals FindMultiModalNeighbors(), RunUMAP(nn.name = "weighted.nn")
  1. Clustering & Annotation
 Group cells and assign biological meaning FindClusters(), FindAllMarkers()
  1. Downstream Analysis
 Discover enriched proteins per cluster, refine cell types FindMarkers(assay = "ADT"), external multimodal workflows

1.10 scRNA-seq vs scRNA-seq + CITE‑seq

Approach Type Purpose
scRNA‑seq Unimodal (RNA only) Characterize transcriptomic heterogeneity: identify cell types, states, and gene expression dynamics based solely on RNA.
CITE‑seq Multimodal (RNA + proteins) Integrate RNA + surface proteins: refine cell type annotation, validate clusters, and capture functional context beyond transcriptomics.

1.11 References