When Doublets Look Like Double-Positive Cells in Flow Cytometry

Technical Blog Draft | Flow Cytometry | Reader-Refined Edition

A CD3/CD19 double-positive population appears where T and B cells used to separate cleanly. Does it represent an unexpected cell phenotype, or have two cells passed through the cytometer as one event? The fluorescence plot alone cannot tell you.

Flow cytometry counts detected events, not independently verified single cells. Cell aggregates or coincident particles can combine signals and alter reported frequencies, although some cell associations reflect genuine biological interactions. Doublet discrimination helps identify events that warrant exclusion or closer investigation. Understanding what the gates can and cannot establish is as important as drawing them [1–3].

In This Article

1. How Doublets Distort Flow Cytometry Results
2. Understanding FSC-A, FSC-H, and FSC-W
3. Building a Singlet Gating Strategy for Your Sample
4. What Changes After Doublet Exclusion?
5. Why Doublets Form and How to Reduce Them
6. When Singlet Gating Is Not Enough
7. Conclusion

1. How Doublets Distort Flow Cytometry Results

A recorded event can contain more than one cell for different reasons. In a physical aggregate, cells are attached before acquisition. In coincidence, otherwise separate particles pass so closely together that they are not resolved as distinct events. Cell–cell conjugates are another kind of physical association, but may arise from meaningful biological contact rather than sample handling. All three can combine signals from more than one cell [2,3].

Consider a CD3-positive T cell passing the detector alongside a CD19-positive B cell. The combined event may fall in a CD3/CD19 double-positive gate even though neither cell coexpresses both markers. Such events can distort both the apparent double-positive fraction and the number of events in its parent population. They also matter in cell sorting, where a sorter may be expected to deliver one cell per destination [2,3].

That does not make every unexpected mixed-lineage event an artifact. The question is whether the two signals arise from one cell, an unresolved coincidence, or a cell complex; pulse data may narrow those possibilities but cannot always distinguish them.

2. Understanding FSC-A, FSC-H, and FSC-W

A passing particle produces a signal pulse that may be recorded as area (A), height (H), and width (W). Area is the integrated signal, height its peak, and width a measure of pulse duration; the implementation varies by instrument. These parameters may be available for forward scatter (FSC) and side scatter (SSC), if acquired and saved [1,4].

Imagine two similarly sized cells crossing the interrogation region in quick succession. If their signals merge, the pulse may be wider and have a larger area without a proportional increase in height. Plotting FSC-A against FSC-H or FSC-W can therefore expose events that fall outside the main singlet-like distribution. The gate needs to follow the sample and instrument, not a preset ratio.

The geometry is less predictable when cells overlap, differ greatly in size, or have unusual shapes. Some pairs may raise pulse height, while large or activated single cells may look atypical. Rico et al. showed why scatter-only discrimination can become difficult in certain activated or proliferating samples [1]. Figure 1 illustrates a common pattern, not a diagnostic signature.

Figure 1. Conceptual pulse geometry; independently drawn from the principles discussed in Rico et al. [1]. Not experimental data.

3. Building a Singlet Gating Strategy for Your Sample

Begin with the cells you intend to analyze. A broad FSC/SSC gate can reduce obvious debris, but it should not remove relevant small or irregular cells. A viability marker may help exclude dead-cell events from analysis; it cannot break up aggregates already present in the suspension. These steps define the population for subsequent pulse analysis, rather than establishing singlet identity.

Next, examine FSC-A versus FSC-H within that population. Events with disproportionately high area relative to the main singlet-like band are candidates for further assessment. Depending on the instrument and sample, FSC-A versus FSC-W or an SSC pulse comparison may reveal additional unusual events. An extra gate is useful only if it improves discrimination without discarding a meaningful population [1,2].

Build the gates from the observed distributions and suitable controls. Then inspect the excluded events on scatter and marker plots. If many are large, activated, or dividing cells relevant to the study, the exclusion may be too strict. Apply a consistent analytical rationale across samples, while recognizing that real shifts in cell shape can affect how a fixed boundary performs.

Keep the parent-gate definition, pulse plots, excluded-event counts, and acquisition settings with the analysis. They make it possible to revisit a surprising population change rather than treating the final plot as the entire record [4].

Figure 2. Illustrative FSC pulse plots with simulated points. No universal gate boundaries are implied. Based on methods described in Rico et al. [1] and Stadinski & Huseby [2].

4. What Changes After Doublet Exclusion?

Suppose a dual-positive population represents 12% of events before singlet gating and 4% of events within the resulting singlet gate. These are hypothetical values, not a typical or expected reduction. The drop suggests that pulse-excluded events contributed to the original double-positive fraction. It does not identify every excluded event as a doublet, nor establish that the 4% remaining are genuine coexpressing single cells.

First check what each percentage uses as its denominator. Singlet gating can remove several phenotypes, not just dual-positive events; a change from 12% to 4% therefore combines changes in the numerator and the parent population. Compare the underlying event counts and gate hierarchy before attributing the shift to aggregates.

Then look separately at the retained and excluded dual-positive events. Do they cluster near pulse-gate boundaries? Are their scatter features unusual? Do single-stain and fluorescence-minus-one controls support the fluorescence gates? These checks may identify a gating or staining explanation, but none alone reveals whether both markers belong to the same cell [2,3]. Table 1 summarizes the conclusions that remain possible after gating.

Table 1. Interpretive examples; these statements are not diagnostic rules.

5. Why Doublets Form and How to Reduce Them

Many aggregates form before a sample reaches the cytometer. Damaged cells can release material that promotes clumping; tissue dissociation, delays in processing, and natural cell adhesion can add to the problem. Gentle resuspension, sample-appropriate filtration, and carefully chosen dissociation conditions may help, provided they do not preferentially remove or damage the cells being studied [4].

Coincidence has a different origin: two unbound particles reach the interrogation region too close in time to be resolved. Concentration and acquisition rate can affect its frequency, so a controlled dilution or rate comparison may be informative. That test does not, by itself, distinguish pre-existing aggregates from stable biological conjugates.

Whether to disrupt an association depends on the experiment. If the aim is to study T-cell conjugates or platelet–leukocyte interactions, breaking the complexes could erase the biology of interest. In that setting, preserving the association and using a method that can resolve its components may be more useful [2,3].

6. When Singlet Gating Is Not Enough

Some tightly associated cell pairs produce pulses that fall inside an otherwise reasonable singlet gate. Stadinski and Huseby discuss the limitation; Burel et al. documented T-cell-containing complexes among events classified as live singlets in human blood samples [2,3]. Passing a pulse gate means an event has the expected pulse characteristics—not that its single-cell identity has been independently confirmed.

For an unexpected dual-positive population, the next test should address the remaining uncertainty. Back-gating shows where the events sit relative to pulse boundaries. A controlled event-rate comparison can probe coincidence. Single-stain and fluorescence-minus-one controls help investigate spillover and gate placement, including possible tandem-dye compensation changes in panels that use those reagents. If the instrument records appropriate fluorescence pulse parameters, those may supply another view; Rico et al. found them useful in the conditions studied, but they are not a universal substitute for scatter measurements [1].

When the identity of the event remains central to the conclusion, imaging flow cytometry or microscopy of sorted material can help determine whether one or multiple cells are present. Imaging itself requires suitable controls, but it tests a question that conventional pulse plots cannot settle. Simply tightening a singlet gate is unlikely to resolve that uncertainty [2,3].

Conclusion

Pulse-based doublet discrimination makes flow cytometry results easier to interpret, but it is not proof that every retained event represents one cell. FSC area, height, and width can flag suspicious events; preparation and acquisition choices can reduce avoidable aggregation or coincidence; and unexpected double-positive populations sometimes need imaging confirmation. The objective is to understand what each event may represent before treating its fluorescence pattern as a cellular phenotype.

References

1. Rico LG, Bardina J, Bistué-Rovira À, et al. Accurate identification of cell doublet profiles: Comparison of light scattering with fluorescence measurement techniques. Cytometry A. 2023;103(5):447–454. doi:10.1002/cyto.a.24690.

2. Stadinski BD, Huseby ES. How to Prevent Yourself from Seeing Double. Cytometry A. 2020;97(11):1102–1104. doi:10.1002/cyto.a.24045.

3. Burel JG, Pomaznoy M, Lindestam Arlehamn CS, et al. The Challenge of Distinguishing Cell–Cell Complexes from Singlet Cells in Non-Imaging Flow Cytometry and Single-Cell Sorting. Cytometry A. 2020;97(11):1127–1135. doi:10.1002/cyto.a.24027.

4. Cossarizza A, Chang HD, Radbruch A, et al. Guidelines for the use of flow cytometry and cell sorting in immunological studies (third edition). Eur J Immunol. 2021;51(12):2708–3145. doi:10.1002/eji.202170126.