10x Next GEM vs GEM-X vs BD Rhapsody: Choosing a scRNA-seq Workflow

10x Next GEM vs GEM-X vs BD Rhapsody: Choosing a scRNA-seq Workflow

Next GEM, GEM-X and BD WTA workflow comparison

Choosing a single-cell RNA sequencing (scRNA-seq) service requires more detail than "10x or BD." A quotation may describe historical 10x Next GEM 3' v3.1, 10x GEM-X 3' v4, or BD Rhapsody whole transcriptome analysis (WTA). These are three expression workflows worth distinguishing before committing samples. Their instrument compatibility, partitioning, recovery guidance and handling options differ.

This comparison is for researchers, biotech teams and CROs planning gene-level discovery studies. Start with the tissue, target cell population and required evidence. A newer reagent may improve useful information per sample, while an established workflow may help maintain continuity with a historical cohort. Neither a brand name nor a high nominal capture percentage guarantees representative data.

The main comparison is conventional 3' expression profiling. Dedicated receptor assays, fixed-sample probe workflows and targeted panels appear as alternative routes when the research endpoint changes. For outsourcing scope, see Outsourcing scRNA-seq: Project Scope, Deliverables and Quotes.

Understand the Names Before Comparing Quotes

Chromium is a platform; Next GEM and GEM-X are technology generations. The assay and reagent version determine the measurement. BD Rhapsody is another platform whose cartridge, bead chemistry and WTA protocol must likewise be specified. Comparing complete workflows avoids treating unlike product names as equivalent instruments.

LayerExamplesWhat the Customer Should Confirm
Partitioning instrumentChromium Controller; Chromium X/iX; compatible BD Rhapsody instrumentChip or cartridge compatibility, instrument configuration and available QC. These instruments do not sequence finished libraries.
Technology and reagentNext GEM 3' v3.1; GEM-X 3' v4; named BD bead and WTA versionsExact kit, protocol revision and permitted input range.
MeasurementConventional 3' expression; other routes such as 5' or FlexGene coverage and any additional libraries needed for the endpoint.
Downstream sequencingCompatible short-read sequencing of prepared librariesRead structure, indexes, allocated read pairs and pooling.

The cited GEM-X 3' v4 guide supports Chromium X, iX and Xo; historical Next GEM guides include Controller workflows. Compatibility is assay-specific. An instrument upgrade alone does not establish that the project uses GEM-X, and changing the sequencer does not convert an end-tag library into full-length isoform data. Record the complete configuration in the Statement of Work.

Next GEM 3' v3.1 vs GEM-X 3' v4

Both conventional workflows capture polyadenylated transcripts and generate 3'-biased gene-expression libraries. They support clustering, annotation and appropriately designed differential-expression analysis. Neither provides every RNA class, complete transcript structures or reliable paired immune receptors by default. The comparison concerns usable gene-level information and project execution, not a change from expression profiling to another modality.

ComparisonNext GEM 3' v3.1GEM-X 3' v4Practical Interpretation
Guide-listed scale500–10,000 cells/sample500–20,000 cells/sampleRanges describe the specified workflows, not guaranteed retained singlets.
Illustrative loading tableAbout 1,650 loaded for 1,000 recoveredAbout 1,450 loaded for 1,000 recoveredManufacturer planning entries; preparation losses and biological QC are additional.
Approximate multiplets at 1,000 recovered0.8%0.4%From the cited loading tables. Rates rise with loading.
Approximate multiplets at 10,000 recovered8%4%Do not extrapolate linearly beyond the validated configuration.
Expression sensitivityHistorical v3.1 baselineManufacturer reports improved sensitivityValidate the relevant population and genes at comparable depth; no fixed improvement applies to every sample.

These figures come from CG000315 Rev F and CG000731 Rev B [5,6]. The manufacturer also advertises recovery up to 80% with GEM-X technology [7]; this maximum is not identical to the conservative loading-table ratio or QC-passing yield. Increased target capacity does not remove doublets or make more cells a substitute for independent donors.

When to Evaluate the Newer Workflow

For a new conventional expression study, GEM-X 3' v4 is a reasonable configuration to assess against sample feasibility and cost. CD Genomics describes this route in its Chromium X Single-Cell and Single-Nucleus RNA-Seq Service. Ask whether any gain changes your intended endpoint: target-gene detection, rare-population coverage or usable cells per specimen. Median genes across all cells can conceal poor recovery of a particular lineage.

For an ongoing v3.1 cohort, first check whether continued use is available and scientifically appropriate. If migrating, include shared bridge specimens where feasible, balance biological groups across versions and document processing differences. A change in detected populations or expression may reflect preparation, chemistry or analysis. Computational integration cannot reliably separate biology from chemistry when they are completely confounded.

10x 3' Gene Expression vs BD Rhapsody WTA

The central cross-platform comparison is GEM-X 3' v4 versus a specified BD WTA workflow. Older Next GEM comparisons provide context, not a direct current-generation ranking. The BD Rhapsody scWTA Service is an expression option; confirm the actual cartridge, bead version and analysis pipeline proposed for your samples.

Input-to-retained-cell comparison across RNA workflowsFigure 2. Illustrative comparison of suspension review, partitioning, library preparation and QC-passing expression profiles. Preparation losses, multiplets and filtering are separate gates. No observed yield or guaranteed platform ranking is implied.

Partitioning and Sample Handling

Chromium co-partitions cells and gel beads into droplets, where conventional expression barcoding occurs. Clumps, debris and unsuitable viscosity can interfere with partitioning. BD captures cells in microwells, binds transcripts to capture beads and retrieves beads for subsequent processing. It avoids the same droplet-channel clogging mechanism, but washing and retrieval can introduce losses. Both workflows remain vulnerable to damaged specimens and ambient RNA.

BD optical QC is available with a compatible Scanner. It estimates occupancy, viability and cell–bead pairing before sequencing, helping identify loading problems. These observations do not identify every molecular doublet or guarantee target-cell representation. For Chromium, combine suspension checks, partitioning observations where available, cDNA assessment and sequencing QC. Neither an instrument trace nor a visually occupied well establishes a successful biological measurement.

Recovery, Multiplets and Low Input

Separate cells received, cells loaded, observed cell-associated barcodes and QC-passing singlets. Compare target-lineage counts at each stage. A specimen with fewer than 2,000 viable cells may lose substantial material during sorting or transfers before reaching either instrument. Evaluate protocol input ranges and amplification requirements rather than declaring microwells universally better for scarce samples.

The 2023 BD HT brochure reports approximately 80% optical capture for a specified cell mixture and multiplets of 2.5% at 10,000 loaded cells [8]. The 10x tables above use recovered-cell denominators. These are different configurations and measurements, not evidence that one brand has a universally lower doublet rate. Both have loading-dependent multiplets; microwells do not physically exclude every second cell. Sample labels detect some cross-sample multiplets, not all same-sample doublets.

Sensitivity, Target Populations and Background

Compare gene and molecular-barcode recovery within the cell type central to the hypothesis, using comparable sequencing allocations. Low-RNA granulocytes may be disproportionately affected by capture, cell calling and filtering. Stromal populations may respond differently to dissociation and handling. A result favorable for immune profiling cannot automatically select a workflow for fibrosis, brain or epithelial atlases.

Ambient RNA can make damaged cells resemble viable populations or create misleading marker expression. Review raw and filtered matrices, empty-partition profiles and contamination estimates. Cleanup can reduce contamination while removing fragile target cells. Better handling can add more value than deeper sequencing. The Single Cell Suspension Preparation Service can support preparation assessment before the final configuration is chosen.

What Published Benchmarks Can—and Cannot—Tell You

Salcher et al., 2024: paired human prostate tumor and benign tissues were compared with Next GEM 3' v3.1 and BD WTA. The reported fractions passing QC were 54.6% and 55.3%, respectively. These describe retention among sequenced cells, not input-to-output capture. Cell-type and gene-specific differences matter more than treating these close percentages as a platform verdict [1].

Colino-Sanguino et al., 2024: the main complex-tissue comparison used mouse mammary tumors, including experimentally damaged samples; human granulocyte-containing specimens were also evaluated. Stromal representation and granulocyte sensitivity differed. This supports testing relevant populations, while mouse-tumor results do not establish the same ranking in every human tissue [2].

De Simone et al., 2025: nine commercial kits were evaluated using PBMCs from one donor. The study helps distinguish usable read allocation, gene sensitivity and cost. Its findings should not be generalized to every tissue or used to rank current versions that were not tested [3]. Manufacturer performance claims and peer-reviewed matched comparisons are different kinds of evidence.

Choose by Sample and Research Endpoint

Project SituationWorkflows to EvaluateDecision RationaleCheck Before Scaling
New discovery atlas from viable cellsGEM-X 3' v4 and BD WTACompare relevant lineages and useful expression information.Target-cell representation, background and comparable read depth.
Continuing a historical v3.1 cohortExisting Next GEM workflow if available; bridged GEM-X transitionPreserve interpretability when reagent generation changes.Group balance, bridge specimens and preprocessing consistency.
Granulocyte-rich blood or inflamed tissueInclude BD WTA and current GEM-X in a pilotHistorical benchmarks suggest population-specific differences.Independent input composition, low-RNA cell calling and retained singlets.
Scarce sorted cells or small research biopsyExact low-input configuration on either platformHandling and preparation may dominate nominal capture efficiency.Available input, transfer loss and permitted amplification range.
Frozen tissue with poor whole-cell recoveryCompatible nuclei workflowNuclei can improve feasibility but change transcript and population representation.Nuclear integrity, target-lineage yield and assay-specific compatibility.

For frozen or difficult tissues, use Nuclei Isolation or Tissue Dissociation? to assess preparation choices. Cell viability thresholds should not be copied directly into nuclei QC. Likewise, neither workflow restores spatial coordinates after tissue dissociation; expression associations require independent validation when used to explain a mechanism.

Selecting Next GEM, GEM-X or BD WTA by project needsFigure 3. Decision framework based on specimen feasibility, cohort continuity and target-cell evidence. Conditional branches lead to a pilot or an alternative assay when conventional expression profiling cannot meet the endpoint.

Compare Total Project Cost and Deliverables

Request configurations with the same biological objective and comparable output scope. In a hypothetical six-sample study targeting 5,000 retained cells/sample at 30,000 read pairs/cell, the expression allocation is 900 million read pairs. This is arithmetic, not a price quotation or claim that six samples provide adequate power. Background, unequal library pooling and recovery changes can alter the required raw allocation.

Compare preparation, capture units, libraries, sequencing, analysis and contingencies. Ask how insufficient input, failed libraries, extra preparation and additional sequencing are handled. A lower capture fee may not lower cost per usable target-cell profile. Conversely, greater cell capacity adds little value if the study needs more donors or the relevant population is lost upstream.

  • Base outputs: Agreed FASTQ files, checksums, unfiltered and filtered count matrices, sequencing summaries and per-sample QC. Record references, software versions and cell-calling criteria.
  • Biological outputs: Specify annotation, donor-aware comparisons and other interpretation separately. Advanced trajectory analysis is not implied by a basic expression matrix.
  • Comparable metrics: Compare target singlets, usable expression counts and population-specific coverage; use the same denominator and state exclusions.

For allocation decisions, consult scRNA-seq Budget Planning. For analysis scope, review the Single-Cell RNA-Seq Data Analysis Service.

When Another Workflow Is More Appropriate

If paired TCR/BCR linked to expression is essential, evaluate compatible 10x 5' receptor enrichment or BD WTA with a dedicated receptor assay. Conventional 3' expression counts do not provide that output by default. See 3' vs 5' scRNA-seq for the library-level decision.

Fixed or FFPE-derived specimens may favor an appropriate probe-based Flex route, with species, gene coverage and preparation assessed separately. A defined gene set may suit BD targeted expression profiling, but unmeasured genes cannot be discovered. Chromatin, surface-tag and guide readouts need additional compatible assays; they are not substitutes for the conventional expression comparison.

What to Validate Before Scaling Up

A hypothetical granulocyte project should prioritize low-RNA representation; a stromal atlas should prioritize fibroblast and endothelial coverage. Use representative pilot specimens, comparable allocations and agreed decision gates. Review preparation loss, usable singlets, target-gene detection, ambient RNA and population bias together. Confirm key findings independently, using suitable input phenotyping or orthogonal assays. Record when repeating preparation, changing assay or narrowing the endpoint is preferable to more sequencing. The scRNA-seq Data Validation Guide provides further quality considerations.

For a fair pilot, define the comparison before inspecting results. Match sample preservation and processing time, document any enrichment, and compare both total and target-cell recovery. Downsample sequencing when interpreting detection differences. More genes at higher depth is not evidence of higher chemistry sensitivity. Assess markers and population coverage alongside molecular-barcode counts; do not select a workflow simply because its filtering retains more barcodes.

Technical aliquots help compare handling, but do not replace biological replication. For treatment comparisons, define the donor or animal as the independent unit and use an appropriate donor-aware model. A pilot may establish preparation feasibility without estimating cohort-level power precisely. Report exclusions and software versions, and revisit decisions when target-cell coverage differs substantially between specimens. Practical QC and analysis guidance should be adapted to tissue context rather than reduced to universal mitochondrial or gene-count thresholds [4].

FAQs

Plan a Workflow Around Your Research Endpoint

CD Genomics supports research-use-only project planning, sample preparation, expression profiling and bioinformatics. Explore Single-Cell Sequencing Services with your specimen details and required results to develop a version-specific project configuration.

References

  1. Salcher S, Heidegger I, Untergasser G, et al. Comparative analysis of 10X Chromium vs. BD Rhapsody whole transcriptome single-cell sequencing technologies in complex human tissues. Heliyon. 2024;10(7):e28358.
  2. Colino-Sanguino Y, Rodriguez de la Fuente L, Gloss B, et al. Performance comparison of high throughput single-cell RNA-Seq platforms in complex tissues. Heliyon. 2024;10(17):e37185.
  3. De Simone M, Hoover J, Lau J, et al. A comprehensive analysis framework for evaluating commercial single-cell RNA sequencing technologies. Nucleic Acids Research. 2025;53(2):gkae1186.
  4. Heumos L, Schaar AC, Lance C, et al. Best practices for single-cell analysis across modalities. Nature Reviews Genetics. 2023;24(8):550–572.
  5. 10x Genomics. Chromium Next GEM Single Cell 3' Reagent Kits v3.1 (Dual Index) User Guide. CG000315, Rev F. Accessed October 12, 2026.
  6. 10x Genomics. Chromium GEM-X Single Cell 3' v4 Gene Expression User Guide. CG000731, Rev B. Accessed October 12, 2026.
  7. 10x Genomics. The Chromium Advantage: Single cell with confidence. Manufacturer technology comparison; performance statements require their stated conditions. Accessed October 12, 2026.
  8. BD Biosciences. BD Rhapsody HT Single-Cell Analysis System. BD-80577, v1.0. 2023. Manufacturer demonstrations. Accessed October 12, 2026.
  9. BD Biosciences. BD Rhapsody System mRNA Whole Transcriptome Analysis (WTA) and Sample Tag Library Preparation Protocol. Protocol for the specified WTA and sample-tag workflow. Accessed October 12, 2026.
For research use only, not intended for any clinical use.

Online Inquiry

For research purposes only, not intended for clinical diagnosis, treatment, or individual health assessments.