GAGE-seq vs ATAC + RNA Multiome: Which Regulatory Layer Does Your Study Need?

Inquiry

Deciphering how non-coding genomic elements govern cell-type-specific gene expression is a central challenge in modern regulatory genomics. Single-cell RNA sequencing (scRNA-seq) profiles transcriptional output, but gene expression alone does not identify the upstream chromatin states or three-dimensional genome features associated with that output. Multimodal single-cell technologies add regulatory layers to the same-cell measurement, allowing researchers to test more specific hypotheses about how chromatin state and genome architecture relate to transcription.

Two major multi-omic paradigms can connect the non-coding genome to transcription: chromatin accessibility combined with transcriptome (e.g., 10x Genomics Multiome Single-Cell ATAC + RNA) and three-dimensional (3D) chromatin architecture combined with transcriptome (exemplified by GAGE-seq, genome architecture and gene expression by sequencing). Although both approaches provide single-cell regulatory insight, they measure different biophysical layers. Chromatin accessibility identifies DNA regions that are accessible to regulatory machinery and supports motif-based TF hypotheses, whereas 3D genome conformation measures contact patterns and spatial proximity among distal loci and higher-order nuclear domains.

This guide provides a rigorous, deep-dive comparison between GAGE-seq (3D genome + RNA) and ATAC + RNA multiome profiling. We examine what each assay measures, clarify the critical distinction between direct experimental measurement and computational inference, evaluate multi-scale nuclear features, and present an actionable decision framework for study design.

Quick Strategy Picker: Which Regulatory Layer Does Your Study Need?

• Need to identify accessible candidate cis-regulatory elements, TF motif activity, and chromatin accessibility changes → Single-Cell ATAC + RNA Multiome
• Need to measure chromatin contact patterns, enhancer-promoter proximity, and TAD/compartment reorganization → 3D Genome + RNA Co-Assay (GAGE-seq / scHi-C + RNA)
• Need to construct cell-type regulatory models with high cell throughput and standardized commercial workflows → 10x Multiome ATAC + RNA
• Need to resolve higher-order nuclear architecture changes during lineage commitment or structural chromosomal rearrangements → 3D Genome Analysis / Spatial Nuclear Organization
  Key technical note: chromatin accessibility identifies DNA that is accessible to regulatory machinery, whereas 3D conformation assays provide experimental evidence of spatial proximity between genomic loci. Neither accessibility nor physical proximity alone establishes functional regulation or causality. This guide discusses research-use discovery workflows and study design principles, not clinical diagnostic testing.

For a foundational overview of chromatin profiling technologies, see our guide on Spatial Epigenomics Explained: Spatial ATAC-seq vs Spatial CUT&Tag for Real Tissues.

Direct Comparison: 3D Chromatin Architecture + RNA vs Chromatin Accessibility + RNA

Choosing between 3D genome co-assays and accessibility multiome profiling requires understanding how physical chromosome folding differs from local chromatin accessibility.

Quick Glossary of Epigenomic and 3D Genomic Architecture

• Chromatin Accessibility: A biophysical state where genomic DNA is temporarily unoccupied by nucleosomes or repressive heterochromatin complexes, allowing transcription factors and transcriptional machinery to access sequence motifs.
• Proximity Ligation: A chromosome-conformation strategy in which spatially proximate DNA fragments in crosslinked chromatin are ligated, creating junctions that provide experimental evidence of genomic proximity.
• Topologically Associating Domains (TADs): Sub-megabase genomic regions characterized by high internal chromatin interaction frequencies, separated by boundary elements enriched for CTCF and cohesin complexes.
• A/B Compartments: Megabase-scale nuclear spatial segregation where transcriptionally active, open chromatin associates in Compartment A, while inactive, heterochromatic regions partition into Compartment B.
• Enhancer-Promoter (E-P) Contact: Spatial proximity between a distal cis-regulatory element and a gene promoter. Contact evidence can support a regulatory hypothesis, but functional enhancer activity requires independent validation.
• Co-accessibility: A statistical correlation of accessibility states between pairs of genomic loci across single cells, used computationally to infer putative regulatory interactions in the absence of direct 3D contact data.

Figure 2. Conceptual workflows of single-cell multi-omic modalities: proximity-ligation-based 3D genome profiling paired with RNA (GAGE-seq) versus Tn5-based chromatin accessibility paired with RNA (10x Multiome).

Deep Dive: How GAGE-seq Profiles 3D Chromatin Architecture and Gene Expression

Developed to overcome the limitations of separate single-cell Hi-C and single-cell RNA sequencing, GAGE-seq (Genome Architecture and Gene Expression by Sequencing) enables simultaneous measurement of genome-wide 3D chromatin conformation and transcriptional activity within the exact same individual cell.

1. Joint RNA and Chromatin Preparation

GAGE-seq begins with crosslinked, permeabilized cells or nuclei so that RNA and chromatin information can be preserved in the same cellular unit. In the published workflow:

• RNA is reverse transcribed using biotinylated poly(T) or random-hexamer primers that introduce sequences needed for later cDNA barcoding.
• Crosslinked chromatin is fragmented with the four-cutter restriction enzymes CviQI and MseI, which generate compatible 5′-TA ends for proximity ligation.
• DNA contact products and cDNA are then prepared for matched combinatorial indexing, allowing the 3D genome and transcriptome layers to be linked back to the same cell.

2. High-Throughput Combinatorial Indexing

GAGE-seq uses sequential ligation-mediated combinatorial indexing rather than relying solely on droplet microfluidics:

• Cells or nuclei are distributed across a 96-well plate for the first round of DNA and cDNA barcoding, pooled, and redistributed to a second 96-well plate for the next barcode round.
• The resulting barcode combinations identify matched 3D genome and transcriptome profiles from the same cell. The published strategy can be extended with additional barcoding rounds for larger-scale designs.

3. Dual-Modality Information Extraction

Following cell lysis and library construction, sequencing reveals:

• scHi-C Layer: Published GAGE-seq datasets recovered average chromatin-contact counts ranging from approximately 181,000 to 265,000 per cell after study-specific quality filtering, including intrachromosomal and interchromosomal contacts.
• scRNA Layer: The same published datasets recovered thousands to tens of thousands of molecule-level transcript counts per cell, with the exact yield varying substantially across cell lines, brain tissue, and bone marrow cells.

For research groups investigating higher-order chromatin structure in tissue and cellular contexts, see our Spatial Nuclear Organization Analysis Services.

Deep Dive: How Single-Cell ATAC + RNA Multiome Profiling Works

In contrast to proximity ligation, the 10x Multiome Single-Cell ATAC + Gene Expression platform measures the biochemical accessibility of non-coding elements alongside transcript abundance.

1. Single-Nucleus Isolation and Permeabilization

Because mature cell membranes can impede dual-assay efficiency, the workflow begins with optimized single-nucleus isolation. Intact nuclei are permeabilized, leaving the nuclear membrane intact while allowing enzymatic reagents to enter.

2. In-Nuclei Tn5 Tagmentation

Permeabilized nuclei are incubated with adapter-loaded Tn5 transposase. Tn5 preferentially integrates sequencing adapters into accessible chromatin, enriching for open regulatory DNA such as promoters and candidate enhancers. This provides a direct accessibility readout, while TF motif enrichment and footprint-like patterns are downstream computational inferences rather than direct measurements of TF occupancy.

3. Shared Single-Nucleus Barcoding

Tagmented nuclei are partitioned with barcode-bearing gel beads in a Chromium workflow so that chromatin-accessibility fragments and gene-expression molecules from the same nucleus receive a shared cell identity. This same-nucleus pairing is the main analytical advantage over generating ATAC-seq and RNA-seq in separate cell populations.

4. Library Separation and Parallel Sequencing

Following droplet breakup, the cDNA and tagmented DNA are separated and converted into two independent, high-complexity sequencing libraries:

• ATAC Library: Quantifies open chromatin peak accessibility genome-wide and maps transcription factor binding motifs.
• Gene Expression (GEX) Library: Quantifies gene expression counts linked to the exact same cell barcode.

Explore our dedicated Single-Cell ATAC + RNA Profiling Service and our standalone Single-Cell ATAC Sequencing Services.

Figure 3. Strategic decision framework for choosing between chromatin accessibility + RNA and 3D genome + RNA according to the primary biological question.

Direct Measurement vs Computational Inference: Enhancer-Promoter Links

A fundamental point of confusion in single-cell epigenomics is how enhancer-promoter (E-P) interactions are established in each assay.

The In Silico Co-Accessibility Approach (ATAC + RNA)

When using ATAC + RNA multiome data, E-P links are inferred computationally, not directly observed. Algorithms such as Cicero, ArchR, and Signac calculate the statistical co-variation of open chromatin peaks across single cells:

1. If an enhancer peak and a promoter peak consistently open and close together across the cell population, the algorithm predicts a functional connection.
2. If enhancer accessibility correlates with target gene expression across single cells, the candidate regulatory relationship is strengthened, but the association still requires independent validation before it is interpreted as a functional enhancer-target pair.

Limitation: Correlation does not prove physical contact. Distal enhancers on the same chromosome may open simultaneously due to shared transcription factor dynamics without forming a physical chromatin loop with the target promoter. Furthermore, pre-formed (poised) chromatin loops may already physically connect an enhancer and promoter before the enhancer becomes fully accessible.

The Physical Contact Approach (GAGE-seq / scHi-C + RNA)

3D genome co-assays capture ligation products between genomic fragments that were brought into spatial proximity in crosslinked chromatin:

1. Ligation junctions provide experimental evidence that distal DNA fragments were spatially juxtaposed when the chromatin was fixed.
2. Comparing contact frequencies across matched cell states can support candidate enhancer-promoter contact models and reveal higher-order architectural changes.

Limitation: Individual single-cell contact maps remain sparse. Robust cell-type-specific loop calling commonly requires aggregation across biologically matched cells or specialized statistical modeling, and physical proximity alone does not demonstrate that an enhancer functionally regulates a target gene.

To learn more about computational integration strategies, read Spatial ATAC-seq & scRNA-seq Integration Strategy for Spatial Epigenomics.

What Neither Assay Can Establish Alone

• Physical proximity does not prove enhancer function. A chromatin contact can support a regulatory model, but perturbation or orthogonal functional testing may be needed to establish whether the contact affects transcription.
• Chromatin accessibility does not prove TF occupancy. Motif enrichment and footprinting generate TF-binding hypotheses; direct protein-DNA occupancy requires an orthogonal assay when that distinction matters.
• Cross-modality correlation does not establish causality. Associations between chromatin state and RNA expression should be interpreted as regulatory hypotheses unless supported by temporal, perturbational, or functional evidence.
• Neither assay replaces biological replication. Replicate design, batch-aware analysis, and independent validation remain important for generalizing findings beyond the profiled cells.

What Biological Questions Change Your Choice?

Selecting the optimal assay depends on the specific regulatory mechanism you need to interrogate:

Question 1: "Which transcription factors drive cell differentiation, and which enhancers are active?"

• Optimal Method: 10x Multiome Single-Cell ATAC + RNA
• Why: ATAC-seq directly measures chromatin accessibility. Motif enrichment, footprint-like patterns, and paired TF expression can then be integrated to build gene-regulatory hypotheses across large numbers of single cells. Direct TF occupancy or causal enhancer function requires orthogonal validation.

Question 2: "Does chromatin loop rewiring precede or follow transcriptional activation?"

• Optimal Method: GAGE-seq (3D Genome + RNA)
• Why: In a staged developmental series or time-course design, GAGE-seq can compare 3D genome architecture and transcription in the same cells at matched biological stages. Published hematopoietic data show that structural and transcriptional changes can be discordant; however, a single cross-sectional measurement cannot by itself establish whether a structural change precedes or follows transcription.

Question 3: "Are disease-associated non-coding variants altering enhancer function or disrupting TAD boundaries?"

• Evaluating the Mechanism:
  • If the variant is hypothesized to alter local accessibility or TF motif activity at an enhancer: use ATAC + RNA Multiome, followed by direct occupancy or perturbation assays when functional proof is required.
  • If the variant is hypothesized to disrupt a boundary or alter long-range regulatory contacts: use 3D Genome Analysis / Hi-C, with DNA-level structural validation where a genomic rearrangement is being claimed.

Question 4: "We need approximately 10,000–20,000 nuclei per sample across a larger multi-sample study."

• Optimal Method: 10x Multiome Single-Cell ATAC + RNA
• Why: Commercially standardized microfluidic platforms provide defined workflows, established computational pipelines, and current configurations supporting up to approximately 20,000 nuclei per sample, making them practical for larger multi-sample studies (10x Genomics Chromium Single-Cell RNA-Seq).

Technical Feasibility and Study-Design Considerations

Key Quality Control Checkpoints for Epigenomic Multiome Studies

Robust multimodal single-cell studies benefit from predefined QC checkpoints, but numerical thresholds should be interpreted in the context of assay version, sample type, library complexity, and run-level controls:

1. Nuclear Isolation and Membrane Integrity
   For both ATAC+RNA and 3D genome co-assays, intact nuclei are essential. Check nuclear morphology under fluorescence microscopy using DAPI and Trypan Blue. Ensure minimal blebbing, minimal clumping, and absence of cytoplasmic debris.
2. Tagmentation and Insert Size Distribution (ATAC Layer)
   High-quality single-cell ATAC libraries must display clear nucleosomal periodicity on a Bioanalyzer or TapeStation, showing distinct peaks corresponding to nucleosome-free fragments (<100 bp), mononucleosomes (~200 bp), and dinucleosomes (~400 bp).
3. Cis-to-Trans Interaction Balance (3D Genome Layer)
   Track the relative abundance of intrachromosomal and interchromosomal contacts together with duplication, mapping quality, and contact-distance distributions. An unexpectedly high random-trans fraction can indicate compromised nuclei or background, but no single cis-to-trans ratio should be treated as a universal GAGE-seq pass threshold.
4. TSS Enrichment and Peak-Overlap Metrics
   For ATAC data, evaluate TSS enrichment and the fraction of high-quality fragments overlapping accessible peaks. TSS enrichment values in the mid-to-high single digits are commonly observed in well-performing datasets, while peak-overlap fractions can vary substantially by tissue, chemistry, and analysis pipeline. Interpret both against assay-specific guidance rather than a single universal cutoff.

To explore single-cell ATAC principles in depth, see What is Single-Cell ATAC Sequencing?.

Common Pitfalls and Practical Tips

• Do not treat co-accessibility as a substitute for physical contact maps.
  Co-accessibility links generated by tools like Cicero provide valuable hypothesis-generating associations, but they must not be reported as confirmed physical chromatin loops without orthogonal 3D validation.
• Do not over-lyse nuclei during permeabilization.
  Excessive lysis detergent concentrations compromise nuclear envelope integrity, leading to leakage of nuclear RNA and cross-contamination of tagmented genomic DNA across droplets.
• Avoid under-sequencing the 3D genome library.
  Because 3D contact matrices sample pairwise combinations across the entire genome, low sequencing depth will result in severe contact dropout, limiting your analysis to coarse A/B compartments rather than fine-scale enhancer-promoter loops.
• Check whether one modality dominates joint clustering.
  Frameworks such as Seurat WNN can learn modality-specific information weights per cell, but the resulting clusters should still be inspected against ATAC and RNA separately to confirm that technical depth or one modality is not dominating the biological classification.

Data and Code Traceability

For 10x Multiome datasets, primary processing commonly uses Cell Ranger ARC, followed by downstream analysis in Signac/Seurat or ArchR. The published GAGE-seq workflow used barcode-aware demultiplexing, BWA for DNA alignment, STAR for RNA alignment, and pairtools for chromatin-contact parsing, together with custom processing for multiway contacts and single-cell integration. Always document software versions, genome assemblies, barcode rules, filtering parameters, and reference annotations.

FAQs

Ready to Map Epigenomic Regulation in Your Research

Connecting non-coding chromatin regulation to transcriptional output can add mechanistic context across developmental biology, oncology, neuroscience, and immunology. The most informative design depends on whether the primary question concerns local accessibility, higher-order genome architecture, or both, and whether the evidence needed is associative, structural, or functional. GAGE-seq is discussed here as an emerging research technology for scientific comparison rather than as a CD Genomics service offering.
For research teams whose question is best addressed by paired chromatin accessibility and gene-expression profiling, CD Genomics provides research-use-only project consultation, nuclei quality assessment, paired chromatin accessibility and gene-expression library construction, and bioinformatics support; explore our Single-Cell ATAC + RNA-seq Service.

References

1. Zhou T, Zhang R, Jia D, et al. GAGE-seq concurrently profiles multiscale 3D genome organization and gene expression in single cells. Nature Genetics. 2024;56(8):1701–1711.
2. Bai D, Zhu C. Single-cell technologies for multimodal omics measurements. Frontiers in Systems Biology. 2023;3:1155990.
3. Liu R, Xu R, Yan S, et al. Hi-C, a chromatin 3D structure technique advancing the functional genomics of immune cells. Frontiers in Genetics. 2024;15:1377238.
4. Fang K, Wang J, Liu L, Jin VX. Mapping nucleosome and chromatin architectures: A survey of computational methods. Computational and Structural Biotechnology Journal. 2022;20:3955–3962.
5. 10x Genomics. Sequencing Requirements for Single Cell Multiome ATAC + Gene Expression. Official technical documentation; updated 2024.