CITE-seq vs. 10x Multiome: Choosing the Right Single-Cell Multi-Omics Platform for Your Research Question

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-20

Key Takeaways

• CITE-seq simultaneously measures cell surface protein expression via antibody-derived tags (ADTs) and RNA transcripts from viable single cells, making it ideal for immune phenotyping and defining cell states based on canonical surface markers like CD antigens.
• 10x Multiome profiles chromatin accessibility (ATAC) and gene expression from isolated nuclei, enabling the study of regulatory genome dynamics, enhancer activity, and developmental trajectories, and is compatible with frozen tissue samples.
• CITE-seq protein detection is generally more sensitive than RNA detection for low-abundance targets due to direct antibody binding, while 10x Multiome's nuclear RNA capture can lead to reduced detection of cytoplasmic-enriched transcripts.
• Cost considerations differ significantly: CITE-seq involves upfront antibody panel costs that are more economical at larger cell numbers, whereas 10x Multiome incurs higher sequencing costs due to dual-library generation (ATAC + RNA) per nucleus.
• Computational analysis for CITE-seq involves integrating separate RNA and ADT count matrices, while 10x Multiome requires processing RNA, ATAC fragment files, and peak count matrices, with integration strategies focusing on paired measurements within the same cell or nucleus.
• Sample requirements are a critical differentiator: CITE-seq necessitates viable single-cell suspensions, limiting its use to fresh or cryopreserved samples that dissociate well, whereas 10x Multiome's nuclei-based approach offers broader compatibility with fresh and frozen tissues, including those challenging for single-cell dissociation.

Researchers planning a single-cell study face a critical decision when selecting between CITE-seq and 10x Multiome platforms. CITE-seq simultaneously measures surface protein expression and RNA transcripts within the same cell, while 10x Multiome captures chromatin accessibility and gene expression from the same nucleus. The choice between these platforms determines which biological questions can be answered, the sensitivity of each molecular measurement, the total cost per cell, and the computational strategies required for data integration and interpretation. This article provides a structured comparison of both platforms across molecular layers measured, technical sensitivity, cost considerations, and downstream analysis possibilities, with decision criteria grounded in specific research objectives.

Understanding the Molecular Layers Each Platform Measures

CITE-seq, which stands for Cellular Indexing of Transcriptomes and Epitopes by Sequencing, pairs antibody-derived tags (ADTs) with single-cell RNA sequencing. The platform uses oligonucleotide-conjugated antibodies that bind to cell surface proteins. After washing away unbound antibodies, the oligonucleotide tags are captured alongside mRNA transcripts during the standard single-cell library preparation. The resulting sequencing library contains both cDNA from expressed genes and antibody-derived tags that quantify surface protein abundance. This design directly links protein-level measurements with transcriptional activity in the same cell.

10x Multiome, formally known as the Chromium Single Cell Multiome ATAC + Gene Expression assay, simultaneously profiles chromatin accessibility and gene expression from individual nuclei. The assay combines assay for transposase-accessible chromatin with sequencing (ATAC-seq) and RNA sequencing within a single workflow. Transposase enzymes fragment open chromatin regions and tag them with sequencing adapters, while reverse transcription captures mRNA from the same nucleus. The resulting libraries provide paired measurements of regulatory DNA accessibility and transcriptional output.

The fundamental distinction lies in what each platform interrogates. CITE-seq measures the functional products of gene expression at the protein level, capturing post-transcriptional regulation, protein stability, and surface phenotypes that define cell states. 10x Multiome measures the regulatory landscape of the genome, revealing which enhancers and promoters are accessible for transcription. Protein expression and chromatin accessibility represent different regulatory layers, and the choice between them depends on which biological process is under investigation.

For immunology studies focused on cell surface markers, CITE-seq provides direct measurement of cluster of differentiation (CD) markers, cytokine receptors, and checkpoint proteins. These protein measurements often correlate imperfectly with mRNA levels due to post-transcriptional regulation, protein degradation rates, and translational control. Researchers studying immune cell subsets defined by canonical surface markers will find CITE-seq more directly aligned with their experimental framework.

For developmental biology and gene regulation studies, 10x Multiome offers direct access to the regulatory genome. Chromatin accessibility changes often precede transcriptional changes during cell fate commitment, providing earlier markers of differentiation trajectories. Researchers investigating enhancer activity, transcription factor binding sites, and regulatory element dynamics will benefit from the chromatin accessibility measurements that 10x Multiome provides.

Technical Sensitivity and Data Quality Considerations

The sensitivity of each platform differs substantially across its respective molecular layers. CITE-seq protein measurements typically achieve high sensitivity for surface proteins because antibody binding is specific and the oligonucleotide tags are amplified efficiently during library preparation. The number of detectable proteins depends on the antibody panel size, with commercial panels ranging from dozens to hundreds of antibodies. Protein detection is generally more robust than mRNA detection for low-abundance targets because antibodies bind directly to their epitopes without requiring transcription or translation.

RNA sensitivity in CITE-seq experiments depends on the underlying single-cell RNA sequencing chemistry. Most CITE-seq protocols use 3-prime capture approaches, which generate sequencing libraries from the 3-prime ends of transcripts. This approach provides reliable gene expression quantification but offers limited isoform resolution compared to full-length sequencing methods. The RNA capture efficiency follows the same limitations as standard single-cell RNA sequencing, with dropout events for lowly expressed genes being a common challenge.

10x Multiome RNA sensitivity is generally lower than standard single-cell RNA sequencing because the assay operates on nuclei instead of whole cells. Nuclear RNA represents a fraction of total cellular RNA, with cytoplasmic transcripts being excluded. This limitation particularly affects detection of genes whose transcripts are predominantly cytoplasmic, including many immediate early response genes and certain metabolic enzymes. Researchers should expect reduced detection of cytoplasmic-enriched transcripts when using 10x Multiome.

Chromatin accessibility sensitivity in 10x Multiome depends on sequencing depth and the number of accessible regions per nucleus. The assay captures fragments from open chromatin regions genome-wide, with typical experiments detecting tens of thousands of peaks per nucleus. However, the sensitivity for individual regulatory elements varies with cell type and chromatin state. Heterochromatic regions and closed enhancers produce few or no fragments, limiting the ability to assess regulatory elements that are inactive in the sampled cell population.

Quality control metrics differ between platforms. For CITE-seq, key metrics include the number of unique molecular identifiers (UMIs) per cell for both RNA and antibody-derived tags, the percentage of mitochondrial reads, and the fraction of reads mapping to antibody tags versus mRNA. For 10x Multiome, additional metrics include the fraction of fragments in peaks (FRiP), the transcription start site (TSS) enrichment score, and the ratio of nucleosome-bound to nucleosome-free fragments. These metrics guide filtering decisions during data preprocessing.

At a Glance: Platform Comparison Table

Cost Structure and Experimental Scale

The cost per cell differs between platforms, and the total experiment cost depends on the number of cells profiled, the sequencing depth required, and the complexity of the antibody panel. CITE-seq requires purchasing oligonucleotide-conjugated antibodies, which represent a significant upfront cost. Commercial panels with hundreds of antibodies cost substantially more than smaller custom panels. The antibody cost is incurred per experiment regardless of the number of cells captured, making CITE-seq more cost-efficient at larger cell numbers.

10x Multiome does not require antibody reagents, but the dual-library nature of the assay increases sequencing costs. Each nucleus generates both an ATAC-seq library and an RNA library, and both require adequate sequencing depth for meaningful analysis. The total sequencing cost per nucleus is typically higher than for standard single-cell RNA sequencing because two molecular layers must be sequenced. Researchers should budget for approximately double the sequencing reads per cell compared to RNA-only assays.

Library preparation costs also differ. CITE-seq library preparation follows standard single-cell RNA sequencing protocols with an additional step to amplify and purify antibody-derived tags. The antibody tag library is typically sequenced separately or pooled with the cDNA library at controlled ratios. 10x Multiome requires separate library preparation for ATAC fragments and cDNA, with distinct amplification and purification steps for each library type.

The choice of sequencing platform affects cost and data quality. CITE-seq libraries require paired-end sequencing with read lengths sufficient to capture the antibody tag barcode and the RNA transcript sequence. 10x Multiome libraries require sequencing configurations that accommodate both the short ATAC fragments and the longer cDNA fragments. Most core facilities and commercial providers offer standardized sequencing configurations for both assay types.

For researchers working with limited budgets, the per-sample cost of 10x Multiome may be prohibitive for large cohort studies. CITE-seq with a focused antibody panel targeting specific cell populations can be more economical when the research question centers on known surface markers. Conversely, studies requiring genome-wide regulatory profiling will find 10x Multiome necessary despite the higher cost.

Sample Requirements and Tissue Compatibility

The input material requirements differ fundamentally between the two platforms. CITE-seq requires viable single-cell suspensions because antibody staining occurs on live cells before fixation and library preparation. This requirement limits CITE-seq to fresh or cryopreserved samples that can be dissociated into single cells without compromising surface epitope integrity. Enzymatic dissociation protocols must be optimized to preserve surface proteins, as prolonged digestion can cleave epitopes and reduce antibody binding.

10x Multiome operates on isolated nuclei, which can be prepared from fresh or frozen tissue. The ability to use frozen tissue is a major advantage for clinical samples, biobanked specimens, and tissues that are difficult to dissociate into viable single cells. Nuclei isolation protocols are compatible with snap-frozen tissue, allowing researchers to profile samples collected under conditions that would be incompatible with CITE-seq.

Tissue types that are challenging for single-cell dissociation, such as adipose tissue, bone, and certain solid tumors, may be more tractable with 10x Multiome because nuclei isolation bypasses the need for viable cell recovery. However, nuclei isolation introduces its own challenges, including the loss of cytoplasmic RNA and potential contamination from ambient nuclear RNA. Researchers should validate that their tissue of interest yields high-quality nuclei with acceptable RNA detection.

CITE-seq is better suited for blood, immune cells, and other samples that naturally exist as single cells or dissociate readily. Peripheral blood mononuclear cells, cultured cell lines, and lymphoid tissues are standard CITE-seq applications. The requirement for viable cells also enables downstream functional assays, such as sorting cells of interest after CITE-seq analysis for expansion or functional testing.

For spatial context, neither platform directly provides spatial information. However, the choice of platform affects compatibility with spatial validation approaches. Protein measurements from CITE-seq can be validated using immunohistochemistry or immunofluorescence on adjacent tissue sections. Chromatin accessibility measurements from 10x Multiome can be validated using ATAC-seq on bulk tissue or spatial chromatin assays where available.

Computational Analysis Workflows

The bioinformatics pipelines for CITE-seq and 10x Multiome share common steps but diverge in modality-specific processing. Both platforms generate raw sequencing data that require demultiplexing, alignment, and count matrix generation. The Bioconductor project provides extensive packages for single-cell analysis, including workflows for multimodal data processing and integration. Researchers should establish reproducible analysis pipelines using version-controlled scripts and containerized environments.

CITE-seq data processing involves separate count matrices for RNA and antibody-derived tags. The RNA count matrix is generated by aligning reads to the transcriptome and counting UMIs per gene. The antibody tag count matrix is generated by identifying antibody barcodes and counting their associated UMIs. Quality control for antibody tags includes assessing the fraction of cells with detectable signal for each antibody, identifying background binding, and normalizing across antibodies to account for differences in binding efficiency.

10x Multiome data processing generates three primary outputs: the RNA count matrix, the ATAC fragment file, and the peak count matrix. The ATAC data requires alignment to the genome, identification of accessible chromatin regions, and quantification of fragments per peak per nucleus. Peak calling can be performed on aggregated data across all nuclei or on individual cell populations after clustering. The Galaxy Training Network offers accessible tutorials for single-cell ATAC-seq analysis that are applicable to 10x Multiome data.

Data integration across modalities is a central computational challenge for both platforms. For CITE-seq, the RNA and protein measurements are inherently paired within each cell, allowing joint analysis without cross-modal alignment. Methods such as weighted nearest neighbor analysis integrate RNA and protein data by constructing a shared cell-cell similarity graph. For 10x Multiome, the RNA and ATAC measurements are also paired within each nucleus, enabling similar joint analysis approaches.

Cross-sample integration is required when combining data from multiple experimental batches or conditions. The nf-core documentation describes community-developed pipelines for single-cell analysis that include batch correction and data harmonization steps. Recent computational methods have advanced the capacity for integrating weakly linked modalities, where the correlation between molecular layers is modest. One study describes a hypergraph contrastive learning framework called MMIHCL that models cell relationships for multimodal data integration, demonstrating high-quality integration across weakly linked datasets and accurate cross-modality feature prediction (MMIHCL study).

Cell type annotation requires different strategies for each platform. CITE-seq provides direct protein measurements of canonical surface markers, enabling annotation based on established immunophenotyping criteria. The protein data can be used to validate RNA-based clustering and to identify cell populations that are transcriptionally similar but phenotypically distinct. For 10x Multiome, cell type annotation often relies on transferring labels from reference RNA datasets using computational methods designed for cross-modality label transfer. One framework called CellPredX uses domain adaptation and deep metric learning to transfer cell type labels between scRNA-seq and scATAC-seq data, with an interpreter module that identifies key markers driving predictions (CellPredX study).

Integration with Other Data Types

Both platforms can be extended to measure additional molecular layers beyond their core assays. CITE-seq can be combined with T cell receptor (TCR) sequencing to profile immune cell clonotypes alongside surface protein expression and transcriptional state. Recent methodological advances have simplified VDJ profiling from 3-prime directed workflows, including single-cell and single-nucleus RNA sequencing approaches. One study describes circVDJ-seq, a method for cost-efficient TCR profiling from 3-prime directed workflows including ATAC + RNA multi-omics, enabling immune repertoire analysis in diverse clinical contexts (circVDJ-seq study).

10x Multiome can be extended with additional assays that measure other regulatory layers. The ATAC component provides chromatin accessibility, which can be integrated with transcription factor motif analysis to infer regulatory networks. Computational methods that model gene expression and chromatin accessibility on a shared time scale enable RNA velocity inference from multi-omic data. One study presents MultiVeloVAE, a probabilistic framework for multi-sample RNA velocity inference that integrates single-cell RNA and multi-omic data, modeling gene expression and chromatin accessibility on a shared time scale and enabling statistical testing of velocity parameters (MultiVeloVAE study).

The choice of platform affects the feasibility of integrating with other data types. CITE-seq protein data integrates naturally with flow cytometry and mass cytometry datasets, enabling cross-validation of cell populations across platforms. 10x Multiome chromatin data integrates with bulk ATAC-seq, ChIP-seq, and other epigenomic datasets, providing context for regulatory element activity. Researchers should consider which external datasets they plan to integrate with their single-cell results.

Spatial transcriptomics integration is possible with both platforms, though the approaches differ. CITE-seq protein markers can be used to map cell populations onto spatial transcriptomics data using computational deconvolution or label transfer. 10x Multiome chromatin accessibility data can be integrated with spatial ATAC-seq or used to infer regulatory activity in spatial contexts. The choice of platform should consider the availability of complementary spatial datasets for the tissue of interest.

Decision Framework Based on Research Questions

The selection between CITE-seq and 10x Multiome should be driven by the primary biological question, with secondary considerations including sample availability, budget, and computational resources. The following decision criteria provide a structured approach to platform selection.

For studies focused on immune cell phenotyping, CITE-seq is the preferred platform when the research question centers on surface protein expression. Immune cell subsets are canonically defined by combinations of surface markers, and CITE-seq provides direct measurement of these markers alongside transcriptional state. Studies of immune activation, exhaustion, and differentiation that rely on surface phenotypes will benefit from CITE-seq protein data.

For studies of gene regulation and chromatin dynamics, 10x Multiome is the preferred platform when the research question centers on regulatory element activity. Developmental trajectories, cellular reprogramming, and disease-associated chromatin changes are directly interrogated by chromatin accessibility measurements. The paired RNA and ATAC data enable linking regulatory element activity to transcriptional output.

For studies requiring both protein and chromatin measurements, neither platform provides both layers. Researchers may need to choose which molecular layer is more informative for their specific question or consider running both assays on matched samples. The integration of CITE-seq and 10x Multiome data from the same tissue requires computational alignment across modalities, which is an active area of method development.

Sample availability often constrains platform choice. Fresh viable cells are required for CITE-seq, while frozen tissue is compatible with 10x Multiome. Clinical studies with limited access to fresh tissue may be forced to use 10x Multiome. Conversely, studies with established protocols for viable cell isolation may prefer CITE-seq for its protein measurements.

Budget considerations may also determine platform choice. The total cost per cell, including reagents and sequencing, should be estimated for both platforms using current pricing from core facilities or commercial providers. For studies requiring large cell numbers, the per-cell cost difference becomes a major factor. For studies with modest cell numbers, the upfront antibody cost for CITE-seq may be acceptable.

Practical Implementation Steps

Implementing either platform requires careful experimental design and quality control. The following steps provide a structured approach to planning and executing a CITE-seq or 10x Multiome experiment.

First, define the biological question and identify the molecular layers that are most informative. If surface protein phenotypes are central to the question, select CITE-seq. If chromatin regulatory dynamics are central, select 10x Multiome. Document the specific markers or regulatory elements that will be measured and how they relate to the expected cell populations.

Second, assess sample availability and compatibility. For CITE-seq, confirm that viable single-cell suspensions can be prepared from the tissue of interest. For 10x Multiome, confirm that nuclei isolation protocols are established for the tissue type. Test sample preparation protocols on pilot samples before committing to the full experiment.

Third, design the antibody panel for CITE-seq or the sequencing depth strategy for 10x Multiome. For CITE-seq, select antibodies that cover the surface markers relevant to the expected cell populations, including positive and negative markers for each population. For 10x Multiome, determine the sequencing depth required for both RNA and ATAC libraries based on the expected cell complexity and the number of cells to be profiled.

Fourth, establish quality control thresholds before data collection. Define the minimum number of RNA UMIs, the maximum mitochondrial read fraction, and the minimum number of detected genes per cell for CITE-seq. For 10x Multiome, add thresholds for TSS enrichment, fraction of fragments in peaks, and nucleosome binding ratio. These thresholds should be based on published guidelines and pilot data.

Fifth, plan the computational analysis pipeline. Select software tools for alignment, count matrix generation, quality control, normalization, clustering, and visualization. Establish version control for analysis scripts and document software versions for reproducibility. The EMBL-EBI Training portal offers learning pathways for single-cell analysis that cover both platforms.

Sixth, allocate resources for data storage and computation. Single-cell multi-omics datasets are large, with raw sequencing data, aligned files, and count matrices requiring substantial storage. Computational requirements for alignment and clustering may exceed local computing capacity, necessitating cloud resources or high-performance computing clusters.

Records and Measurements for Quality Assurance

Maintaining detailed records of experimental parameters and quality metrics is essential for reproducible single-cell multi-omics research. The following measurements should be recorded for every experiment.

For sample preparation, record the tissue source, dissociation protocol, cell viability before and after processing, and the time between sample collection and library preparation. For CITE-seq, record the antibody panel composition, antibody concentrations, staining conditions, and washing steps. For 10x Multiome, record the nuclei isolation protocol, nuclei concentration, and the ratio of nuclei to transposition reaction components.

For library preparation, record the number of cells or nuclei loaded, the capture efficiency, and the resulting library concentrations. Record the PCR amplification cycles for each library type and any deviations from the standard protocol. For CITE-seq, record the ratio of antibody tag library to cDNA library in the sequencing pool. For 10x Multiome, record the sequencing configuration for both ATAC and RNA libraries.

For sequencing, record the sequencing platform, read length, sequencing depth per cell, and the number of reads allocated to each library type. Monitor sequencing quality metrics including per-base quality scores, duplication rates, and alignment rates. These metrics provide early indicators of library quality and guide decisions about resequencing or data filtering.

For data analysis, record the software versions, reference genome versions, and parameter settings for each analysis step. Document the quality control thresholds applied and the number of cells or nuclei retained after filtering. Record the clustering parameters and the number of clusters identified. These records enable reproducibility and facilitate troubleshooting when results are unexpected.

The Carpentries lessons provide foundational training in data management and reproducible analysis practices that are directly applicable to single-cell multi-omics projects. Adopting these practices early in a project reduces the risk of data loss and analysis errors.

Common Failure Patterns and Troubleshooting

Several recurring problems affect CITE-seq and 10x Multiome experiments. Recognizing these failure patterns early enables corrective action before resources are wasted.

Low cell recovery is a common failure in both platforms. For CITE-seq, low recovery often results from cell loss during antibody staining and washing steps. For 10x Multiome, low recovery may result from nuclei loss during isolation or clumping that blocks microfluidic channels. Monitoring cell counts at each step and optimizing centrifugation and resuspension conditions can improve recovery.

High background signal in antibody tags is a specific CITE-seq failure mode. This problem arises when unbound antibodies are not completely washed away or when antibodies bind nonspecifically to dead cells or debris. The fraction of reads mapping to antibody tags versus mRNA should be monitored, and excessive antibody background may require optimization of washing conditions or antibody concentrations.

Low chromatin accessibility signal in 10x Multiome can result from suboptimal transposition conditions, excessive nuclei fixation, or degradation of open chromatin during sample preparation. The TSS enrichment score and fraction of fragments in peaks provide quantitative measures of chromatin accessibility quality. Low values indicate that the ATAC component of the assay has failed and the data should not be used for regulatory analysis.

Batch effects are a persistent challenge in both platforms. Differences in sample processing, library preparation, and sequencing runs introduce technical variation that can obscure biological differences. Including replicate samples across batches and using computational batch correction methods can mitigate these effects. The nf-core documentation describes pipeline configurations that support batch-aware analysis.

Ambient RNA contamination affects RNA measurements in both platforms. Free-floating mRNA from lysed cells is captured during library preparation, contributing background signal that is particularly problematic for highly expressed genes. Computational methods that estimate and subtract ambient RNA profiles can improve data quality, but prevention through careful sample handling is preferable.

Limitations and Interpretation Boundaries

Both platforms have inherent limitations that constrain biological interpretation. Researchers must understand these boundaries to avoid overinterpreting their data.

CITE-seq protein measurements reflect antibody binding, which depends on epitope accessibility and antibody specificity. Antibodies may cross-react with related proteins or fail to bind when epitopes are masked by protein interactions. The number of proteins measured is limited by the antibody panel, and proteins without validated antibodies cannot be profiled. Protein quantification is relative instead of absolute, and comparisons across experiments require careful normalization.

10x Multiome chromatin accessibility measurements reflect the aggregate accessibility of a population of nuclei, not the binding of specific transcription factors. Open chromatin regions may be accessible without being functionally active, and closed regions may become accessible only under specific conditions. The resolution of ATAC-seq does not identify which transcription factors are bound, requiring computational motif analysis to infer potential regulators.

RNA measurements in both platforms are limited by capture efficiency and sequencing depth. Lowly expressed genes are frequently missed, creating dropout events that complicate differential expression analysis. The 3-prime capture approach used in most CITE-seq protocols limits isoform resolution. Nuclear RNA in 10x Multiome excludes cytoplasmic transcripts, biasing detection toward nuclear-localized RNAs.

Data integration across modalities and samples remains challenging despite methodological advances. The correlation between protein expression and mRNA levels is often modest, and the correlation between chromatin accessibility and gene expression is similarly imperfect. Computational methods that integrate these layers must model the complex relationships between molecular measurements. Recent methods such as scGALA use graph-based learning to improve cell alignment across datasets, identifying more high-confidence alignments without compromising accuracy (scGALA study).

Cell type annotation is inherently limited by the reference data used for label transfer. When reference datasets lack certain cell populations or when the query data contains novel cell states, annotation accuracy decreases. Methods that model continuous phenotypic variation, such as the Φ-Space framework, offer alternatives to discrete label assignment by characterizing cell identity in a low-dimensional phenotype space defined by reference phenotypes (Φ-Space study).

Safety and Regulatory Context

Single-cell multi-omics research involving human samples is subject to ethical and regulatory requirements that vary by jurisdiction. Researchers must obtain appropriate institutional review board approval for human subjects research, including consent for genomic analysis and data sharing. The NCBI Data Resources provide guidance on data submission and access policies for human genomic data, including controlled access for sensitive data.

Data privacy is a critical consideration for single-cell genomics. Single-cell datasets can contain information that identifies individuals, particularly when combined with clinical metadata. Researchers should implement data de-identification procedures, restrict access to raw data, and follow institutional data security policies. Controlled access repositories provide a mechanism for sharing data while protecting participant privacy.

For animal research, institutional animal care and use committee approval is required before sample collection. The choice of platform may affect the number of animals needed, as CITE-seq requires viable cells and may necessitate fresh tissue collection while 10x Multiome can use frozen tissue. Researchers should design experiments to minimize animal numbers while achieving statistical power.

Biosafety considerations apply to sample handling and library preparation. Human samples may contain bloodborne pathogens, requiring appropriate personal protective equipment and biosafety cabinet use. Chemical hazards in library preparation kits, including fixatives and organic solvents, require proper handling and disposal. Institutional biosafety committees can provide guidance on safe practices.

Data sharing policies vary by funding agency and journal requirements. Many funders require deposition of single-cell data in public repositories such as the NCBI Gene Expression Omnibus or the EMBL-EBI ArrayExpress. Researchers should plan for data deposition early in the project and ensure that consent forms permit data sharing in accordance with repository policies.

Professional Escalation Criteria

Researchers should seek expert consultation when encountering specific challenges that exceed their local expertise. The following situations warrant escalation to core facility staff, bioinformatics specialists, or external consultants.

If cell recovery or viability falls below acceptable thresholds during CITE-seq sample preparation, consult with core facility staff who may have optimized protocols for the specific tissue type. Troubleshooting dissociation and staining conditions may require iterative testing that benefits from experienced guidance.

If chromatin accessibility quality metrics are consistently poor in 10x Multiome experiments, consult with specialists who can assess nuclei isolation protocols and transposition conditions. Poor TSS enrichment or low fraction of fragments in peaks may indicate a systematic protocol issue that requires expert troubleshooting.

If computational analysis produces unexpected results, such as poor clustering or failed integration, consult with bioinformatics specialists who have experience with the specific platform. The Bioconductor support site and community forums provide access to experts who can diagnose analysis issues.

If data integration across modalities or samples fails to converge or produces biologically implausible results, consider consulting with computational biologists who specialize in multi-omics integration. Methods such as Mowgli, which combines integrative nonnegative matrix factorization and optimal transport for paired multi-omics data, may require expert configuration for optimal performance (Mowgli study).

If the research question evolves to require additional molecular layers beyond the chosen platform, consult with experts about the feasibility of extending the assay or integrating with complementary datasets. Adding TCR profiling to CITE-seq or spatial validation to 10x Multiome may require specialized protocols and analysis approaches.

A Practical Decision Framework for Platform Selection Based on Experimental Constraints

Beyond the biological question, the choice between CITE-seq and 10x Multiome often hinges on practical constraints that researchers encounter during experimental planning. This section provides a structured decision framework that weighs sample logistics, budget allocation, and analysis capacity against the molecular information each platform provides. The framework is designed to be applied before committing resources to either platform.

Step 1: Assess Sample Availability and Handling Requirements

The first decision point concerns the physical state of the samples available for the study. CITE-seq requires viable single-cell suspensions because antibody staining must occur on live cells with intact surface epitopes. This requirement eliminates CITE-seq for studies relying on archived frozen tissue, fixed specimens, or samples that cannot be dissociated without compromising cell viability. 10x Multiome operates on isolated nuclei, which can be prepared from fresh or frozen tissue, making it the only option for biobanked specimens and clinical samples collected under conditions that preclude viable cell recovery.

For tissues that are difficult to dissociate, such as adipose tissue, bone, and certain solid tumors, 10x Multiome may be more tractable because nuclei isolation bypasses the need for viable cell recovery. However, researchers should verify that their tissue of interest yields high-quality nuclei with acceptable RNA detection before committing to the platform. Pilot experiments with a small number of samples can establish whether nuclei isolation protocols produce sufficient yield and quality for the full study.

Step 2: Evaluate the Required Molecular Resolution

The second decision point addresses whether the research question requires protein-level or chromatin-level information. CITE-seq provides direct measurement of surface proteins, which is essential for studies where cell states are defined by canonical surface markers, such as immune cell phenotyping, tumor microenvironment characterization, and studies of activation or exhaustion markers. Protein measurements capture post-translational regulation and protein stability that mRNA levels do not reflect, making CITE-seq the preferred platform when surface phenotypes are the primary readout.

10x Multiome provides chromatin accessibility measurements that reveal regulatory element activity. This information is essential for studies of gene regulation, developmental trajectories, enhancer-promoter interactions, and transcription factor dynamics. Chromatin accessibility often changes before transcriptional changes during cell fate commitment, providing earlier markers of differentiation. For studies investigating the regulatory mechanisms underlying cell state transitions, 10x Multiome provides the direct measurement of the regulatory genome that CITE-seq cannot offer.

Step 3: Calculate Total Cost Including Hidden Expenses

The third decision point requires a complete cost estimate that includes reagents, sequencing, and computational resources. CITE-seq requires purchasing oligonucleotide-conjugated antibodies, which represent a significant upfront cost that is incurred regardless of the number of cells captured. This cost structure makes CITE-seq more cost-efficient at larger cell numbers because the antibody expense is amortized across more cells. However, the antibody panel must be validated for specificity, and custom panels require additional optimization time and expense.

10x Multiome does not require antibody reagents, but the dual-library nature of the assay increases sequencing costs. Each nucleus generates both an ATAC-seq library and an RNA library, and both require adequate sequencing depth for meaningful analysis. The total sequencing cost per nucleus is typically higher than for standard single-cell RNA sequencing because two molecular layers must be sequenced. Researchers should budget for approximately double the sequencing reads per cell compared to RNA-only assays.

Computational costs also differ between platforms. Both platforms generate large datasets that require substantial storage and processing capacity. 10x Multiome generates additional files, including ATAC fragment files and peak count matrices, which increase storage requirements. The computational complexity of integrating chromatin accessibility with gene expression data may require specialized expertise and additional analysis time. Researchers should factor these costs into their total budget when comparing platforms.

Step 4: Verify Analysis Capacity and Bioinformatics Support

The fourth decision point assesses whether the research team has the computational expertise to process and interpret the data generated by each platform. CITE-seq data processing involves separate count matrices for RNA and antibody-derived tags, requiring familiarity with multimodal analysis workflows. The Bioconductor project provides packages for CITE-seq analysis, including normalization and integration of protein and RNA data. The Galaxy Training Network offers accessible tutorials for single-cell analysis that cover both platforms.

10x Multiome data processing requires additional expertise in ATAC-seq analysis, including alignment to the genome, peak calling, and assessment of chromatin accessibility quality metrics. The analysis pipeline is more complex than CITE-seq because it involves two distinct molecular layers with different data structures. Researchers without prior experience in chromatin analysis should budget time for training or consult with bioinformatics specialists. The EMBL-EBI Training portal offers learning pathways for single-cell analysis that cover both platforms.

Step 5: Consider Downstream Validation Requirements

The fifth decision point considers how the single-cell data will be validated and integrated with other experimental approaches. CITE-seq protein measurements can be validated using flow cytometry or mass cytometry on the same samples, providing orthogonal confirmation of cell populations. The protein markers identified in CITE-seq can also be used for immunohistochemistry or immunofluorescence on tissue sections, linking single-cell data to spatial context.

10x Multiome chromatin accessibility data can be validated using bulk ATAC-seq on matched samples or ChIP-seq for specific transcription factors. The regulatory elements identified in 10x Multiome can be tested using reporter assays or CRISPR perturbation experiments to confirm functional relevance. The choice of platform should consider which validation approaches are feasible and which will provide the most compelling evidence for the biological conclusions.

Applying the Framework to Common Research Scenarios

For an immunology study focused on characterizing T cell subsets in peripheral blood, CITE-seq is the preferred platform when the research question centers on surface markers that define functional states. The viable cell suspension from blood is readily obtained, and the antibody panel can be designed to cover canonical T cell markers. The protein data provides direct measurement of activation and exhaustion markers that may not be reflected in mRNA levels.

For a developmental biology study investigating cell fate commitment in differentiating stem cells, 10x Multiome is the preferred platform when the research question centers on chromatin dynamics during lineage specification. The paired ATAC and RNA data enable linking regulatory element activity to transcriptional output, revealing the sequence of molecular events during differentiation. The ability to use frozen samples also facilitates time-course experiments where samples are collected at multiple time points.

For a clinical study using archived tumor specimens, 10x Multiome is the only viable option when fresh tissue is unavailable. The ability to profile frozen tissue enables analysis of biobanked samples with associated clinical outcomes. The chromatin accessibility data can reveal regulatory programs associated with treatment response or resistance, providing insights that would be inaccessible with CITE-seq.

Documenting the Decision Process

Researchers should document the rationale for platform selection, including the biological question, sample availability, cost estimates, and analysis capacity. This documentation supports reproducibility and provides context for interpreting results. The decision framework should be revisited if experimental conditions change, such as new sample availability or budget adjustments. Recording the decision criteria also facilitates comparison across studies and helps reviewers understand the experimental design choices.

Frequently Asked Questions

What is the main difference between CITE-seq and 10x Multiome?

CITE-seq measures surface protein expression using antibody-derived tags alongside RNA transcripts from the same cell. 10x Multiome measures chromatin accessibility using ATAC-seq alongside RNA transcripts from the same nucleus. The core difference is the molecular layer paired with gene expression: proteins for CITE-seq and regulatory DNA accessibility for 10x Multiome.

Can I use frozen tissue with CITE-seq?

CITE-seq requires viable single-cell suspensions because antibody staining occurs on live cells before library preparation. Frozen tissue is generally incompatible with CITE-seq because freeze-thaw cycles compromise cell viability and surface epitope integrity. 10x Multiome is compatible with frozen tissue because the assay operates on isolated nuclei.

Which platform is more expensive per cell?

The cost comparison depends on the antibody panel size for CITE-seq and the sequencing depth for 10x Multiome. CITE-seq requires purchasing oligonucleotide-conjugated antibodies, which add upfront reagent costs. 10x Multiome requires sequencing both ATAC and RNA libraries, increasing sequencing costs per nucleus. Researchers should estimate total costs for their specific experimental design.

How do I choose between protein and chromatin measurements?

The choice depends on the biological question. If the study focuses on cell surface phenotypes, immune cell subsets, or protein-level regulation, CITE-seq provides direct measurements of the relevant markers. If the study focuses on gene regulation, enhancer activity, or developmental trajectories, 10x Multiome provides direct measurements of chromatin accessibility.

Can I integrate CITE-seq and 10x Multiome data from the same samples?

Integration across platforms is possible but computationally challenging because the molecular layers are different. Methods for cross-modality integration can align cells based on shared RNA measurements, but the protein and chromatin data cannot be directly compared. The MMIHCL framework demonstrates integration of weakly linked modalities using hypergraph contrastive learning.

What quality control metrics should I report for each platform?

For CITE-seq, report the number of RNA UMIs per cell, the number of antibody tag UMIs per cell, the percentage of mitochondrial reads, and the number of detected genes per cell. For 10x Multiome, additionally report the transcription start site enrichment score, the fraction of fragments in peaks, and the nucleosome binding ratio.

Does 10x Multiome work with single cells or only nuclei?

10x Multiome is designed for isolated nuclei, not whole cells. The assay requires nuclear isolation because the transposition reaction and RNA capture are optimized for nuclear preparations. This design enables analysis of frozen tissue but limits detection of cytoplasmic RNA.

How many antibodies can I include in a CITE-seq panel?

The number of antibodies depends on the commercial panel or custom design. Panels ranging from dozens to hundreds of antibodies are available. The practical limit is determined by antibody specificity, background binding, and sequencing capacity for antibody-derived tags. Researchers should validate custom panels for specificity before large-scale experiments.

Related Bioinformatics Guides

• RNA-Seq vs Microarray: Choosing the Right Gene Expression Profiling Platform
• Single-Cell vs Single-Nucleus RNA Sequencing: Choosing the Right Approach
• RNA-Seq Alignment: Choosing the Right Tool and Parameters
• Spatial Proteomics vs. Single-Cell Proteomics: Choosing the Right Approach
• Single-Cell Multi-Omics Integration: Methods and Applications

Related Clinical & Scientific Guides

• A Practical Guide to Detecting Antimicrobial Resistance Genes in Shotgun Metagenomic Data
• Computational Immunology: Modeling the Immune System
• How to Set Hard Filters for Germline Variant Calling: A Practical Guide to GATK Best Practices

References and Further Reading

• NCBI Data Resources. National Center for Biotechnology Information.
• EMBL-EBI Training. European Bioinformatics Institute.
• Bioconductor. Bioconductor Project.
• Galaxy Training Network. Galaxy Project.
• nf-core Documentation. nf-core.
• The Carpentries Lessons. The Carpentries.
• Single-cell data integration across weakly linked modalities.. 2026.
• circVDJ-seq for T cell clonotype detection in single-cell and spatial multi-omics.. 2026.
• CellPredX, a computational framework for cross-data type, cross-sample, and cross-protocol cell type annotation through domain adaptation and deep metric learning.. 2026.
• scGALA advances graph link prediction-based cell alignment for comprehensive data integration and harmonization.. 2025.
• Φ -Space: continuous phenotyping of single-cell multi-omics data.. 2025.
• Inferring differential dynamics from multi-lineage, multi-omic, and multi-sample single-cell data with MultiVeloVAE.. 2025.
• Paired single-cell multi-omics data integration with Mowgli. bioRxiv, 2023.

This article is educational and does not replace validated analysis plans, institutional policy, clinical interpretation, or specialist review.