CITE-seq

The core innovation of CITE-seq is its ability to simultaneously capture the whole transcriptome and a panel of cell-surface proteins from the same single cell. This is achieved by using oligonucleotide-conjugated antibodies (often called Antibody-Derived Tags or ADTs). These antibodies bind to specific surface epitopes, and their conjugated DNA barcodes are captured alongside the cellular mRNA during the standard single-cell RNA sequencing library preparation. This provides a direct, quantitative readout of protein abundance that is orthogonal to mRNA expression.

Surface protein markers are the gold standard for immunophenotyping. CITE-seq leverages this by providing quantitative protein-level data that can resolve cellular identities where transcriptomics alone is ambiguous.

• Immune Cell Discrimination: Distinguishes closely related T-cell subsets (e.g., CD4+ naive vs. memory) using established protein markers like CD45RA and CD45RO.
• Transcriptional Ambiguity: Resolves cell clusters that are transcriptionally similar but functionally distinct based on their surface proteome.
• Validation: Provides orthogonal validation for computationally inferred cell-type annotations derived from mRNA signatures.

CITE-seq generates a paired multi-modal dataset for every cell, requiring specialized computational methods for integrated analysis. Weighted Nearest Neighbor (WNN) analysis, implemented in the Seurat v4 framework, is a key technique. It learns the relative utility of each modality for each cell, constructing a unified graph that reflects both transcriptomic and proteomic similarity. This allows for:

• Integrated Clustering: Defining cell states based on a consensus of both data types.
• Multi-Modal Dimensionality Reduction: Visualizing cells in a shared latent space, such as with WNN-UMAP, that captures the full complexity of the data.
• Bridging Modalities: Using protein data to anchor and interpret transcriptional heterogeneity within a known cell type.

A key biological insight from CITE-seq comes from observing discordance between mRNA and protein levels within a single cell. This discordance is not just technical noise; it is a direct readout of post-transcriptional regulation.

• Translation Rate: A high mRNA count with a low protein count suggests translational repression.
• Protein Stability: A low mRNA count with a high protein count indicates a stable protein with a long half-life, persisting after its transcript is degraded.
• Regulatory Dynamics: Tracking these differences across a developmental trajectory (pseudotime) reveals the temporal dynamics of gene regulation, showing when a gene is transcribed versus when its protein product is functionally required.

Rare cell populations, such as circulating tumor cells or antigen-specific T-cells, can be masked by transcriptional noise in standard scRNA-seq. CITE-seq's protein readout provides a high-fidelity, low-noise signal that dramatically improves the detection of these rare populations. By gating on a specific combination of surface markers (e.g., a tetramer-positive, activated T-cell), researchers can isolate a tiny, functionally critical population and then deeply characterize its full transcriptomic state, revealing its activation program, metabolic state, and clonotype in a way that would be impossible with either modality alone.

The technology is commercially available through Total-Seq™ reagents from BioLegend, offering a standardized and scalable path for adoption. Key features include:

• Pre-Titrated Panels: Ready-to-use panels optimized for human or mouse immune profiling, covering hundreds of canonical markers.
• Custom Conjugation: The ability to conjugate the DNA barcode to any user-specified antibody, enabling the design of bespoke panels targeting specific signaling proteins, activation states, or disease-relevant epitopes.
• Sample Multiplexing: Using distinct lipid- or cholesterol-tagged oligonucleotides (hashtags) to label and pool multiple samples into a single sequencing run, reducing technical batch effects and cost.

The core reagent enabling CITE-seq. Oligonucleotide-conjugated antibodies carry a unique DNA barcode instead of a fluorophore. After binding to cell-surface epitopes, these barcodes are amplified and sequenced alongside the transcriptome. Each ADT's barcode count serves as a quantitative proxy for the abundance of its target protein, allowing simultaneous capture of surface proteomics and whole transcriptomics from the same single cell.

CITE-seq is a quintessential multi-omics integration technology, natively generating paired transcriptome and surface proteome data. Computational methods like weighted nearest neighbor (WNN) analysis in Seurat v4 learn a joint multimodal representation. This allows for:

• Defining cell states based on both mRNA and protein
• Identifying populations invisible to transcriptomics alone
• Mapping post-translational regulation where protein levels diverge from RNA expression

CITE-seq belongs to a broader class of feature barcoding assays. Beyond surface proteins, the same oligonucleotide conjugation strategy can label:

• Hashtag antibodies for sample multiplexing (cell hashing)
• Peptide-MHC multimers for antigen-specific T-cell detection
• CRISPR guide RNAs for pooled genetic screens This unified barcoding framework allows diverse molecular features to be read out on a standard sequencing platform alongside the transcriptome.

Large pre-trained models like scGPT and Geneformer are increasingly adapted for CITE-seq data. These single-cell foundation models can:

• Impute missing protein modalities from transcriptomic data alone
• Generate joint embeddings of RNA and ADT counts
• Transfer annotations across CITE-seq experiments By pre-training on massive multi-modal atlases, these models learn generalizable cellular representations that accelerate downstream analysis of CITE-seq datasets.

Multiple commercial and academic platforms implement oligonucleotide-conjugated antibody profiling:

• CITE-seq (BioLegend/New York Genome Center): The foundational method using poly(dA) tailed antibodies
• REAP-seq: A parallel development using similar antibody-oligo conjugates
• BD AbSeq (BD Biosciences): A commercialized version integrated with the Rhapsody microwell platform All share the same principle but differ in antibody panels, barcode design, and compatible sequencing workflows.

The commercial reagent line from BioLegend that standardizes CITE-seq experiments. TotalSeq antibodies are available in multiple formats:

• TotalSeq-A: Compatible with 10x Genomics 3' single-cell workflows
• TotalSeq-B: Optimized for 10x Genomics 5' immune profiling
• TotalSeq-C: Universal format for plate-based protocols Each antibody is rigorously validated for specificity and carries a unique, amplifiable DNA barcode, enabling reproducible protein quantification across laboratories.