Single-cell ATAC sequencing (scATAC-seq) is a technique that maps which stretches of DNA are physically open and accessible in individual cells, one cell at a time. Where traditional methods average this information across millions of cells, scATAC-seq reveals how the genome’s packaging differs from one cell to the next, even within the same tissue. The technology has become a go-to tool for understanding why cells with identical DNA behave so differently, and it has already reshaped research in cancer, immunology, and developmental biology.
Why “Open Chromatin” Matters
Every cell in your body carries the same DNA, yet a neuron looks and acts nothing like a liver cell. The difference largely comes down to which genes each cell can actually read. DNA is not floating loosely inside the nucleus; it is wound tightly around protein spools called histones and packed into a dense structure called chromatin. When a gene needs to be turned on, the chromatin around it loosens and “opens up,” allowing the molecular machinery to get in and read it. When a region stays tightly wound, the genes there stay silent.
ATAC-seq, which stands for Assay for Transposase-Accessible Chromatin using sequencing, was designed to find those open regions across the entire genome. It can compare which chromatin regions are accessible in different cell types or disease states, reveal where regulatory proteins called transcription factors are binding, and identify previously unknown regulatory elements that control gene activity.1Theoretical and Natural Science. ATAC-seq: A powerful tool for investigating chromatin accessibility and transcription factor binding The “single-cell” version does all of this, but for thousands of individual cells in one experiment rather than a blended average of a whole tissue.
How the Core Chemistry Works
The workhorse of any ATAC-seq experiment is a bacterial enzyme called Tn5 transposase. Tn5 has a useful quirk: it can cut DNA and simultaneously glue short synthetic tags (called sequencing adapters) onto the cut ends. Researchers load Tn5 with these adapters, then let it loose on cells. Because tightly packed chromatin physically blocks the enzyme, Tn5 can only cut and tag the open, accessible portions of the genome. The resulting tagged DNA fragments are then amplified and sequenced, producing a genome-wide map of where chromatin was open.
Biochemical studies have pinpointed where Tn5 attacks DNA wrapped around nucleosomes, finding that its major cleavage sites land near the points where the DNA enters and exits the nucleosome spool.2PubMed Central. Biochemical analysis of nucleosome targeting by Tn5 transposase This precision is part of why ATAC-seq can reveal not just which regions are open, but also the fine-scale positioning of nucleosomes and where transcription factors sit on the DNA.
The entire tagmentation reaction is fast and requires very little starting material compared to older methods for studying chromatin. That efficiency is what made it feasible to scale down to single cells in the first place.
Getting to Single-Cell Resolution
Running ATAC-seq on a bulk tissue sample is like averaging every person’s voice in a crowded room. You get a signal, but you lose the individual speakers. In a tumor, for instance, some cancer cells may have wide-open chromatin around drug-resistance genes while neighboring cells do not. A bulk experiment would blur those differences into a single profile. Single-cell ATAC-seq preserves them.
The fundamental challenge is giving each cell’s DNA fragments a unique molecular barcode so that, after sequencing, you can trace every fragment back to the cell it came from. Two main strategies have emerged to solve this problem.
Combinatorial Indexing
The first widely used approach, called sci-ATAC (single-cell combinatorial indexing ATAC), skips the step of physically isolating individual cells into their own tiny compartments. Instead, it performs the Tn5 tagmentation reaction in many small pools of cells, each pool receiving a distinct barcode built into the Tn5 adapters. The cells are then reshuffled and split into new pools, where a second barcode is added during amplification. The probability that any two cells end up with the same pair of barcodes is vanishingly small, so each unique barcode combination effectively labels a single cell. This approach profiled chromatin accessibility in over 15,000 single cells in its original demonstration and can scale even higher without requiring specialized microfluidic equipment.3PubMed Central. Multiplex single cell profiling of chromatin accessibility by combinatorial cellular indexing
Droplet-Based Methods
The second major strategy borrows from the droplet microfluidics revolution that transformed single-cell RNA sequencing. Here, individual cell nuclei are captured inside microscopic water-in-oil droplets alongside barcoded beads. The Tn5 tagmentation is typically performed before the cells are loaded onto the microfluidic chip, and the barcoding happens inside the droplet. One platform called HyDrop, for example, uses dissolvable hydrogel beads that release uniquely barcoded primers inside each droplet; after heating to deactivate the Tn5 and free the DNA fragments, barcoded copies are made, the emulsion is broken, and all fragments are pooled for sequencing.4eLife. Hydrop enables droplet-based single-cell ATAC-seq and single-cell RNA-seq using dissolvable hydrogel beads In one run on frozen mouse brain tissue, this platform produced nearly 8,000 high-quality single-cell chromatin profiles with strong enrichment at gene-regulatory regions.5eLife. Hydrop enables droplet-based single-cell ATAC-seq and single-cell RNA-seq using dissolvable hydrogel beads
Commercial platforms from companies like 10x Genomics use a similar droplet-based philosophy and have become the most common way labs run scATAC-seq today. Semi-automated benchtop workflows have also appeared that push data quality even higher. One such method benchmarked on standard cell lines achieved median unique fragment counts above 23,000 per cell and transcription start site enrichment scores well above quality thresholds, with the vast majority of input cells passing established quality-control criteria.6Nature Communications. Semi-automated IT-scATAC-seq profiles cell-specific chromatin accessibility in differentiation and peripheral blood populations
The Data Sparsity Problem
Single-cell ATAC-seq data looks nothing like, say, a spreadsheet of gene expression values. You get a massive table of cells versus genomic regions, and the vast majority of entries are zero. Any given cell only has two copies of the genome, so each accessible region can be cut by Tn5 at most a couple of times. Most regions will not be captured at all in any individual cell, purely by chance. The result is what researchers call extreme sparsity: a matrix with thousands to millions of features, almost all of them empty for any one cell.7Genomics, Proteomics & Bioinformatics. Computational Analyses and Challenges of Single-cell ATAC-seq
This is not a flaw in the experiment; it is a basic consequence of the biology. Each cell’s genome offers limited material for the enzyme to find, so coverage per cell is inherently low.8PubMed Central. Computational Analyses and Challenges of Single-cell ATAC-seq Overcoming this sparsity is the central computational challenge of the field, and it has driven the development of specialized analysis software that can pool information across similar cells to fill in the gaps.
Making Sense of the Data
Once you have raw sequencing reads tagged with cell barcodes, the analysis pipeline goes through several stages. Reads are aligned to a reference genome, duplicates are removed, and cells that pass quality filters are kept. The genome is then divided into windows or “peaks” of accessibility, and cells are clustered based on how similar their open-chromatin patterns look. Specialized tools like SnapATAC can then call peaks on each cluster’s aggregated data, identify regulatory elements specific to each cell type, and run motif enrichment analysis to discover which transcription factors are likely driving each group’s chromatin state.9Nature Communications. Comprehensive analysis of single cell ATAC-seq data with SnapATAC
Because open chromatin does not directly tell you which genes are being expressed, researchers often want to connect the accessibility data to gene expression. This is where multi-omic approaches and computational integration become essential.
Pairing Chromatin Access with Gene Expression
Knowing that a stretch of DNA is open does not automatically tell you what gene it controls, or whether that gene is actively being read. The most powerful studies combine scATAC-seq with single-cell RNA sequencing so that, ideally from the same cells, you get both the chromatin landscape and the actual transcription output. Methods have been developed that generate both ATAC-seq and mRNA-seq data simultaneously from the same cells, and the resulting dual-omic profiles show a strong correlation between promoter accessibility and gene expression, matching what you would see if the two assays were run separately.10PubMed Central. A simple and robust method for simultaneous dual-omics profiling with limited numbers of cells
This pairing matters because chromatin accessibility is not a simple on/off switch for genes. A region might be open in preparation for a gene to turn on later, or it might stay open even after the gene is silenced by some other mechanism. By looking at both layers at once, researchers can start to build cause-and-effect models of gene regulation that neither dataset could support alone.
Applications in Cancer
Cancer research has been one of the biggest beneficiaries of scATAC-seq. Tumors are notoriously heterogeneous: even within a single patient’s tumor, different cancer cells can have dramatically different chromatin profiles, and those differences often determine which cells survive treatment. In prostate cancer, combined single-cell ATAC and RNA sequencing of cells treated with the drug enzalutamide revealed pre-existing cell subpopulations with distinct chromatin states that possessed regenerative potential and persisted through treatment, suggesting they seed relapse.11Nature Communications. Single-cell ATAC and RNA sequencing reveal pre-existing and persistent cells associated with prostate cancer relapse
A similar story has emerged in breast cancer. Integrated single-cell analysis of tamoxifen-resistant breast tumors found striking cell-to-cell heterogeneity and identified nine distinct “cancer states” with unique open-chromatin signatures, including states specific to primary tumors, recurrent tumors, and shared between both. The study pinpointed specific transcription factors tied to a core gene signature of 137 genes, along with metabolic pathways that mediate communication between cancer cell subpopulations and the surrounding immune environment.12PubMed Central. Integrated single-cell analysis reveals distinct epigenetic-regulated cancer cell states and a heterogeneity-guided core signature in tamoxifen-resistant breast cancer These are the kinds of distinctions that bulk assays would entirely miss.
Beyond Cancer
The technology has spread well beyond oncology. In autoimmune disease research, a computational framework called SCADS integrates genetic risk data from genome-wide association studies with scATAC-seq profiles to score individual cells for their relevance to diseases like inflammatory bowel disease. Applied to immune cell populations, this approach revealed that even within a single canonical cell type, like CD8-positive T cells, there is marked variation in disease relevance, and it could pinpoint the underlying gene programs and genetic variants responsible.13PubMed Central. Connecting polygenic disease risk to cell states and regulatory programs through single-cell chromatin accessibility
Developmental biology has embraced scATAC-seq as well. A study of mouse embryonic gonads profiled chromatin accessibility across both sexes during the critical window when the gonad commits to becoming either a testis or an ovary. The data showed that individual cell types could be identified purely by their chromatin landscape and that cells could be arranged along developmental trajectories, providing a detailed epigenetic map of how cell fates become progressively restricted.14PubMed Central. The single-cell chromatin landscape in gonadal cell lineage specification
CRISPR Meets Chromatin Accessibility
One of the more inventive extensions of scATAC-seq combines it with CRISPR-based genetic perturbations. In a method called Perturb-ATAC, researchers use CRISPR to knock down or knock out specific genes in a pool of cells, then run scATAC-seq and read out both the CRISPR guide RNA and the chromatin profile from each cell. This lets you directly connect the loss of a particular gene to changes in chromatin accessibility across the genome.15PubMed Central. Coupled single-cell CRISPR screening and epigenomic profiling reveals causal gene regulatory networks
A related approach, CRISPR-sciATAC, scaled this concept up to target over 100 chromatin-related genes in human leukemia cells, generating chromatin data for roughly 30,000 single cells and revealing how the loss of specific chromatin-remodeling enzymes altered accessibility at the binding sites of individual transcription factors.16PubMed Central. Profiling the genetic determinants of chromatin accessibility with scalable single-cell CRISPR screens Combined CRISPR-and-ATAC approaches have already been used to build gene regulatory networks that identify drug resistance mechanisms. In one study, the network analysis pointed to a transcription factor called ZFPM2 as a driver of resistance to the cancer drug dasatinib, a prediction that was confirmed when knocking down ZFPM2 made the resistant cells sensitive to the drug again.17Cell Reports Methods. CRISPR and transcriptomics-assay for transposase-accessible chromatin enables high-content multimodal perturbation screening
Practical Headaches and Workarounds
Running scATAC-seq is not without frustrations. One persistent nuisance is mitochondrial DNA. Mitochondria sit outside the nucleus but contain their own small genome, and their DNA is highly accessible to Tn5. As a result, a large fraction of the sequencing reads can come from mitochondria rather than the nuclear genome, wasting sequencing capacity. Researchers have tested several fixes, including using CRISPR/Cas9 to selectively cut mitochondrial fragments out of the sequencing library. This approach reduced mitochondrial reads substantially without compromising data quality, while an alternative strategy of removing detergent from the cell-lysis step also cut mitochondrial contamination but at the cost of noisier data and fewer peaks.18Scientific Reports. Reducing mitochondrial reads in ATAC-seq using CRISPR/Cas9
Sample preparation is another pain point. Fresh tissue generally works better than frozen, but flash-frozen samples can still yield good results with the right protocols. Nuclear isolation needs to be clean enough to avoid debris that clogs microfluidic channels but gentle enough to preserve chromatin state. Each tissue type tends to require its own optimization, which is part of why scATAC-seq is still more of a specialist technique than single-cell RNA sequencing.
The Technology in Plants and Non-Model Organisms
Most scATAC-seq work has been done in mammals, but the technology is not limited to them. Plant cells pose unique challenges because of their rigid cell walls, so researchers use isolated nuclei rather than whole cells. A study of Arabidopsis root tips showed that single-nucleus ATAC-seq data correlated strongly with bulk ATAC-seq from the same tissue and could identify cell-type-specific accessible sites. When integrated with single-nucleus RNA sequencing, the chromatin accessibility clusters closely mirrored the gene-expression clusters, confirming that the approach captures meaningful biology even in plant systems.19Molecular Plant. Single-Nucleus Transcriptomes and Chromatin Accessibility Characterize Plant Cell Types
Adding Spatial Coordinates
Standard scATAC-seq experiments dissociate tissue into a suspension of single cells or nuclei, which means you lose all information about where each cell sat in the original tissue. A method called sciMAP-ATAC addresses this by performing combinatorial indexing on tiny microbiopsies taken from recorded positions in a tissue section. Each microbiopsy is small enough (about 214 microns on a side) to retain spatial meaning, and the resulting data is equivalent in quality to non-spatial sci-ATAC, with the added benefit that every cell’s chromatin profile is tied to a known location in the tissue.20Nature Communications. Spatially mapped single-cell chromatin accessibility This matters for questions where tissue architecture is central, like understanding how chromatin states change across layers of the brain cortex or from the edge to the interior of a tumor.