Single cell genomics is a set of technologies that measure genetic activity in individual cells rather than in bulk tissue samples. Traditional methods blend thousands or millions of cells together, producing an average that masks the differences between them. Single cell approaches isolate each cell and read its molecular contents separately, revealing which genes are active, which proteins are present, and how each cell differs from its neighbors. The result is a high-resolution picture of biological complexity that bulk methods simply cannot provide.
Why Individual Cells Matter
A tissue sample from your liver, brain, or a tumor contains dozens of distinct cell types mixed together. When you grind that sample up and sequence it as a whole, you get one blended gene expression profile representing the average of every cell in the mix. That average can be useful, but it hides critical details. A small population of drug-resistant cancer cells, for instance, might make up only 1% of a tumor. Their unique genetic signature gets drowned out in bulk data.
Single cell genomics solves this by profiling gene expression at the level of individual cells. A single experiment can capture data from more than 25,000 genes across 10,000 or more cells, making it possible to identify rare cell types, track cells in transitional states, and distinguish between populations that look identical under a microscope but behave very differently.
How Single Cell Sequencing Works
The workflow has five core steps: isolate individual cells, break them open, amplify their genetic material, sequence it, and analyze the data. Each step has its own technical challenges, but the basic logic is straightforward.
Cell Isolation
The first task is physically separating cells from one another. The most popular approach today uses droplet-based microfluidics. A device channels cells through tiny tubes alongside a stream of oil, encapsulating each cell inside its own microscopic droplet. Each droplet acts as a self-contained reaction chamber, functionally equivalent to a well in a lab plate but with a reaction volume roughly a million times smaller. This keeps each cell’s contents isolated and dramatically reduces the risk of contamination between samples.
Other methods exist, including hydrodynamic cell traps and pneumatic membrane valves, but droplet-based systems dominate because they can process thousands of cells quickly and consistently.
Lysis and Library Preparation
Once a cell is isolated in its droplet, it gets broken open (lysed), releasing its RNA. The amount of genetic material in a single cell is far too small to sequence directly, so it must be amplified first. For RNA-based methods, the released messenger RNA is captured using short molecular tags called barcoded oligonucleotides. These tags serve two purposes: they grab the RNA molecules, and they stamp each one with a barcode that identifies which cell it came from. After that, the RNA is converted into DNA copies and amplified so there’s enough material to sequence.
A critical quality control measure at this stage involves unique molecular identifiers, or UMIs. These are short random sequences attached to each original RNA molecule before amplification begins. Because the amplification process makes many copies of each molecule, UMIs let researchers count how many distinct original molecules were present, filtering out the duplication bias that amplification introduces.
Sequencing and Analysis
The amplified material is then run through a high-throughput sequencer, producing millions of short reads. Bioinformatics software matches each read back to its cell of origin using the barcodes applied earlier, reconstructing a gene expression profile for every individual cell in the experiment.
Making Sense of the Data
Single cell datasets are enormous and sparse. Many genes in any given cell register as zeros, not because they’re truly inactive, but because the tiny amount of starting material means some molecules go undetected. These “dropout events” create statistical challenges that standard genomics tools weren’t designed to handle.
Researchers use specialized software ecosystems to work through this. Seurat (built in R) and Scanpy (built in Python) are the two most widely used platforms, and both offer tools for filtering noise, normalizing data, reducing the dataset’s complexity, and clustering cells into groups based on similar gene expression patterns. Visualization techniques like UMAP compress tens of thousands of dimensions into two-dimensional plots where clusters of similar cells appear as distinct islands.
Batch effects are another persistent challenge. When samples are processed on different days or on different platforms, systematic technical differences can creep in and masquerade as biological variation. Dedicated correction tools like Harmony and scMerge help align datasets so that genuine biological signals aren’t confused with artifacts of the experimental setup.
Beyond RNA: Other Single Cell Modalities
Single cell RNA sequencing (scRNA-seq) is the most established technique, but the field has expanded to measure nearly every layer of cellular biology at single cell resolution.
- Single cell ATAC-seq (scATAC-seq) profiles chromatin accessibility, revealing which regions of a cell’s DNA are physically open and available for gene activation. This gives insight into a cell’s regulatory landscape, not just which genes are on, but which ones are poised to turn on.
- CITE-seq measures both gene expression and surface proteins on the same cell simultaneously. This is especially valuable in immunology, where surface proteins define a cell’s identity and function.
- Single cell methylation sequencing detects chemical modifications to DNA that silence genes, adding an epigenetic layer to the picture.
- Single nucleus RNA-seq (snRNA-seq) captures RNA from just the nucleus rather than the whole cell. This is useful for tissues like brain or muscle where intact cells are difficult to isolate, since frozen samples can still yield intact nuclei.
- Single cell proteomics and metabolomics are newer and less mature but aim to measure proteins and small metabolic molecules at the single cell level.
Multi-omic approaches that combine two or more of these measurements on the same cell are increasingly common. Pairing gene expression data with protein levels or chromatin accessibility on a cell-by-cell basis creates a richer, more integrated picture of cell state than any single measurement alone.
Applications in Cancer Research
Cancer has been one of the most transformative areas for single cell genomics. Tumors are not uniform masses of identical cells. They contain cancer cells in different states, immune cells trying to fight or inadvertently helping the tumor, blood vessel cells, and structural support cells. This complex ecosystem is called the tumor microenvironment, and understanding it is critical for predicting how a patient will respond to treatment.
Single cell RNA sequencing has made it possible to build detailed atlases of this microenvironment across multiple cancer types. Researchers have used standardized protocols to dissociate tumors from nine different cancer types and characterize the immune cell subtypes within each one. This work revealed that certain immune cell subtypes tend to cluster together in “hubs.” One hub resembles organized immune structures associated with coordinated immune responses. A second hub includes inflammatory immune cells, antigen-presenting cells, and T cells expressing markers of tumor reactivity, forming what researchers call a “type-1 immunity hub.” These hubs are not just statistically correlated; they’re spatially co-located within the tumor tissue.
This kind of detailed mapping has direct clinical relevance. By profiling which immune cell subtypes are present and how reactive the T cells are before treatment begins, researchers can better predict which patients will respond to immunotherapy drugs like checkpoint inhibitors. The abundance of specific cell subtypes in each hub has been linked to both early and long-term responses to these treatments.
Mapping Development and Cell Fate
Single cell genomics has also reshaped developmental biology. Stem cells differentiate into specialized cell types through a series of intermediate states, and tracking those transitions was historically difficult because the intermediates are rare and fleeting. Bulk sequencing blends them all together, making it nearly impossible to define the precise branching points where one cell type diverges from another.
With scRNA-seq, researchers can capture cells at every stage of differentiation and computationally order them along a trajectory based on progressive changes in gene expression. Tools like Monocle3, Slingshot, and RNA velocity reconstruct these paths, identifying not just the starting and ending cell types but the transient intermediates and the branch points where lineages split. This works without needing to know in advance what those intermediates look like, since the data itself reveals them.
These trajectory maps have been applied to embryonic development, organ formation, and stem cell biology, providing molecular-level detail about how a single fertilized egg gives rise to hundreds of specialized cell types.
Profiling the Immune System
The immune system presents a unique challenge for genomics: T cells and B cells each carry a receptor that is essentially unique to that cell, generated through random genetic rearrangement. These T cell receptors (TCRs) and B cell receptors (BCRs) determine what each immune cell can recognize and attack. Bulk sequencing can detect which receptor sequences are present in a sample, but it can’t tell you which receptor belongs to which cell or what that cell is doing.
Single cell sequencing solves this by reading the receptor sequence and the full gene expression profile from the same cell. This means researchers can identify the exact pairing of receptor chains (the alpha and beta chains of a TCR, or the heavy and light chains of a BCR), which is critical for understanding what a given immune cell actually recognizes. They can then correlate that receptor identity with the cell’s functional state: Is it activated? Exhausted? Proliferating?
This capability has become a powerful tool for studying immune responses to infections, vaccines, autoimmune diseases, and cancer. Metrics like clonotype diversity, which measure how many distinct receptor sequences are present and how evenly they’re distributed, help quantify the breadth and intensity of an immune response at a level of detail that was previously inaccessible.
Current Limitations
Single cell genomics is powerful, but it comes with real constraints. The dropout problem remains significant: because each cell provides so little starting material, many genes that are genuinely active go undetected, creating datasets full of zeros that require careful statistical handling. Standard tools designed for bulk data tend to produce false positives when applied naively to this sparse data.
Cost and throughput are improving rapidly but still limit some applications. Processing tens of thousands of cells is routine now, but profiling every cell in a whole organ remains impractical for most labs. The tissue dissociation step can also introduce bias, since some cell types survive the process better than others, meaning the cells you sequence may not perfectly represent the cells that were in the tissue.
Finally, single cell sequencing captures a snapshot in time. It tells you what a cell looks like at the moment it was isolated, not how it got there or where it’s going. Computational trajectory analysis can infer developmental paths, but these are predictions based on patterns across many cells, not direct observations of a single cell changing over time. Newer lineage tracing methods that introduce heritable genetic barcodes into living cells are beginning to address this gap, letting researchers track actual cell histories rather than inferring them.