Single-Cell and Spatial Transcriptomics

Single-cell transcriptomics measures gene expression in individual cells rather than in bulk tissue, revealing the heterogeneity that averaging across many cells obscures, while spatial transcriptomics retains the position of each measurement within a tissue section. Together they allow the transcriptome to be read cell by cell and, increasingly, in situ, exposing distinct cell types, states, and their organization within tissues.

Definition

Single-cell transcriptomics quantifies RNA transcripts within individual cells, and spatial transcriptomics measures transcript abundance while preserving the two-dimensional location of each measurement within a tissue, so that expression can be analysed per cell and in anatomical context.

Scope

This topic covers how transcriptomes are captured from single cells (cell isolation, barcoding, and sequencing), the computational steps that turn sparse per-cell data into cell-type maps (clustering and dimensionality reduction), and spatial methods that preserve tissue location. It is a methodological reference within transcriptomics and provides no clinical guidance.

Core questions

• How is RNA captured and uniquely labelled from thousands of individual cells in parallel?
• How are sparse single-cell profiles clustered into cell types and states?
• How can spatial methods preserve where in a tissue each transcript was measured?
• What technical artefacts (dropout, doublets, batch effects) must analysis account for?

Key concepts

• Cell isolation and droplet-based capture
• Cell barcoding and unique molecular identifiers
• Sparsity and dropout
• Dimensionality reduction and clustering
• Cell-type and cell-state identification
• Trajectory and pseudotime analysis
• Spatially resolved transcriptomics
• Batch effects and integration

Mechanisms

In single-cell RNA sequencing, individual cells are separated — often by encapsulation in droplets — and each cell's transcripts are tagged with a cell-specific barcode and frequently a unique molecular identifier before pooled sequencing, so that reads can be assigned back to their cell of origin and counted without amplification bias. Because each cell yields little RNA, the resulting matrices are sparse and noisy: not every expressed gene is detected (dropout), and analysis relies on dimensionality reduction and clustering to group cells into types and states, as Zeisel and colleagues did for the mouse brain. Tang and colleagues first demonstrated whole-transcriptome sequencing of a single cell, establishing the approach. Spatial transcriptomics, introduced by Stahl and colleagues, places tissue sections onto arrays of positionally barcoded capture spots so that expression can be mapped back onto tissue architecture, linking molecular profiles to anatomy.

Clinical relevance

Single-cell and spatial transcriptomics are reshaping reference maps of tissues in development, immunology, and oncology by resolving the cellular composition and spatial organization of normal and diseased tissue. As a reference topic this entry explains how cell-resolved expression evidence is generated; it is not a basis for individual diagnostic or treatment decisions.

Evidence & guidelines

Methodological reference points include the first single-cell transcriptome sequencing (Tang and colleagues), large cell-typing studies (Zeisel and colleagues), and the introduction of array-based spatial transcriptomics (Stahl and colleagues). These are methodological references rather than clinical guidelines.

History

Single-cell transcriptomics began in 2009 when Tang and colleagues sequenced the transcriptome of one cell. Through the mid-2010s, droplet-based barcoding scaled the approach to thousands of cells, enabling systematic cell-type discovery such as the brain atlas of Zeisel and colleagues. In 2016, Stahl and colleagues introduced array-based spatial transcriptomics, and subsequent imaging- and sequencing-based spatial methods extended resolution and gene coverage, complementing large cell-atlas initiatives.

Debates

Resolution versus transcriptome coverage in spatial methods

Imaging-based spatial methods can reach near single-cell or subcellular resolution but typically measure a limited, predefined panel of genes, whereas sequencing-based capture methods cover the whole transcriptome at coarser spatial resolution; the trade-off between resolution and breadth remains an active design question.

Key figures

• M. Azim Surani
• Sten Linnarsson
• Joakim Frisen
• Fuchou Tang

Seminal works

• tang-2009
• zeisel-2015
• stahl-2016

Frequently asked questions

Why is single-cell expression data so sparse?

Each cell contains only a small amount of RNA, so many genes that are truly expressed are not captured in any given cell (dropout). This makes per-cell profiles noisy and incomplete, which is why analysis relies on grouping many cells together through clustering.

How does spatial transcriptomics differ from single-cell RNA-seq?

Single-cell RNA-seq dissociates tissue, so cells are profiled individually but lose their original location. Spatial transcriptomics measures expression while preserving where in the tissue each measurement came from, allowing molecular profiles to be placed back into anatomical context.