Key takeaways

• When two similar cells pass the laser together, pulse area roughly doubles and width roughly doubles, while peak height stays close to that of a single cell.
• That asymmetry is the whole basis of doublet discrimination: doublets have high area relative to height, so they fall off the FSC-A versus FSC-H diagonal.
• The opposite statement — that doublets show higher height relative to area — is a common inversion, and gating on it removes the wrong events.
• Standard sequence: FSC-A vs FSC-H first, then FSC-W or SSC-W as a second pass, with fluorescence area versus height where needed.
• Doublets matter most in DNA content work, where a G1 doublet sits exactly where a G2/M single cell should, inflating the apparent G2/M fraction.
• Prevention beats gating: filter immediately before acquisition, avoid over-concentrating, and run at a lower flow rate to reduce coincidence.
• Do not over-gate. Genuinely large single cells — blasts, activated lymphocytes, megakaryocytes — sit near the doublet region and can be discarded by an aggressive gate.

Contents

• Why doublets matter
• Pulse geometry: area, height and width
• The common inversion
• Step-by-step gating
• Fluorescence-based discrimination
• A worked gating sequence
• Preventing doublets at the bench
• The risk of over-gating
• Validation and tips
• Choosing reagents
• Frequently asked questions

Reagents where doublets distort the result

Doublet exclusion matters most in the assays below — DNA content, apoptosis quadrants and proliferation are all directly corrupted by uncorrected doublets.

Propidium iodide

Propidium Iodide Staining Solution

DNA stainFlow cytometry

DNA content analysis is the application most damaged by doublets — a G1 pair mimics a G2/M cell exactly.

View product →

7-AAD

7-AAD Viability Staining Solution

Viability dyeFlow cytometry

Dead cells and debris promote aggregation, so viability gating and doublet gating work together.

View product →

Annexin V

Annexin V-FITC / 7-AAD Apoptosis Kit

Apoptosis kitFlow cytometry

Doublets of a live and an apoptotic cell land in the double-positive quadrant and inflate late apoptosis.

View product →

EdU

EdU Cell Proliferation Flow Cytometry Assay Kit

ProliferationFlow cytometry

S-phase quantification depends on clean singlet gating to avoid counting aggregates as cycling cells.

View product →

CD3-FITC

FITC Anti-Human CD3 Antibody [OKT-3]

FITCHuman

In immunophenotyping, a doublet of two differently stained cells reads as a spurious double-positive.

View product →

Isotype

FITC Mouse IgG1, Isotype Control [MOPC-21]

FITCMouse IgG1

A matched isotype control separates non-specific staining from the aggregation artefacts above.

View product →

Why doublets matter

A doublet is two cells passing the interrogation point together. The instrument has no way to know it saw two objects; it records one event with the combined signal. The consequences differ by application:

• DNA content analysis is the worst affected. Two G1 cells together carry 2n + 2n = 4n DNA — exactly the content of a genuine G2/M cell. Without doublet exclusion the G2/M fraction is inflated and any cell cycle conclusion is unsafe.
• Immunophenotyping — a doublet of a CD3-positive and a CD19-positive cell reads as a single CD3/CD19 double-positive event, generating a population that does not exist.
• Apoptosis assays — a live cell stuck to an apoptotic one lands in the double-positive quadrant, overstating late apoptosis.
• Rare event detection — when the population of interest is a small fraction of total events, doublet artefacts can rival or exceed the real signal.
• Sorting — a doublet sorted as a target cell delivers an unwanted passenger, which matters for single-cell downstream work.

Pulse geometry: area, height and width

Everything follows from what the detector records as a particle crosses the beam. The signal over time forms a pulse, and the instrument reports three measures of it.

The reason height does not double is worth stating explicitly, because it is the crux. Two cells travelling end to end enter the beam one after the other. At any instant the laser illuminates roughly one cell’s worth of material, so the peak is similar to a single cell. What changes is how long the signal lasts — and therefore the total area beneath it.

So a doublet is characterised by a high area-to-height ratio and a wide pulse. Both are directly exploitable.

The common inversion

An earlier version of this guide stated the rationale the other way round — that doublets show a higher peak height relative to area. That is inverted, and it matters because it points the gate at the wrong events.

To be unambiguous: on an FSC-A versus FSC-H plot with area on the x-axis and height on the y-axis, single cells form a tight diagonal because area and height rise together. Doublets have extra area without extra height, so they sit to the right of, and below, the singlet diagonal. The singlet gate is drawn around the diagonal itself; events displaced toward higher area are excluded.

The same logic in reverse also holds: nothing normal produces high height with low area, so that region is generally empty apart from noise.

Step-by-step gating

1. FSC-A versus FSC-H

The primary and usually sufficient gate. Single cells form a linear population along the diagonal; doublets and aggregates deviate toward higher area. Draw the gate around the diagonal population.

2. FSC-W or FSC-H versus FSC-W

A second pass using pulse width. Singlets show a narrow, tight width distribution; doublets show a distinctly broader one. This catches pairs that the area-height plot missed, particularly cells of unequal size.

3. SSC-A versus SSC-H

The same principle applied to side scatter. Useful in heterogeneous samples where granularity differs between populations, and as a cross-check when forward scatter alone leaves ambiguity.

Two passes are conventional. Applying all three is reasonable for cell cycle work or sorting, where a doublet is especially costly.

Fluorescence-based discrimination

The same pulse geometry applies to fluorescence channels, and it is the standard approach for DNA content analysis.

Plotting PI-A against PI-H separates singlets on a diagonal from doublets displaced toward higher area — for exactly the reason above. Two G1 cells give double the integrated fluorescence but a similar peak, so they sit off the line while a true G2/M cell, with genuinely twice the DNA in one nucleus, sits on it.

That distinction is the entire value of the method: area alone cannot separate a G1 doublet from a G2/M singlet, but area against height can.

A worked gating sequence

Order matters. Doublet exclusion belongs early — before viability and marker gating — because an aggregate carrying a dead cell will otherwise contaminate every downstream gate.

Preventing doublets at the bench

Gating removes doublets from the analysis; it does not recover the cells lost to them. Reducing formation in the first place is more efficient.

• Filter immediately before acquisition through a 35–70 µm mesh. This is the single most effective step, and it must be done just before running rather than earlier.
• Do not over-concentrate. Higher cell density raises the probability of coincidence at the interrogation point.
• Run at a lower flow rate. Slower sample delivery narrows the core stream and reduces the chance of two cells arriving together — the standard trade-off against acquisition time.
• Include EDTA in the buffer where the protocol allows. Chelating calcium reduces cadherin-mediated adhesion.
• Handle gently. Over-vigorous pipetting damages cells, and DNA released from lysed cells is highly effective at gluing the rest together.
• Add DNase for samples with substantial cell death, which addresses that free-DNA problem directly.

The risk of over-gating

A tight singlet gate is not automatically better. Large single cells legitimately produce larger pulses, and can sit close to the doublet region — blasts, activated lymphocytes, megakaryocytes and many tumour cells among them.

An aggressive gate drawn on a lymphocyte-dominated sample will quietly discard exactly the population of interest in a leukaemia or activation study. The safeguard is to check what the excluded events are before accepting the gate, rather than tightening until the plot looks neat.

It is also worth recording the singlet percentage. A sudden drop between samples usually signals a preparation problem rather than biology.

Validation and tips

• Check against a known single-cell control — a well-dissociated cell line — to confirm the gate sits where it should.
• Re-optimise per cell type. Gates drawn for lymphocytes will not suit adherent lines or primary tissue digests.
• Use more than one parameter. Area-height with a width plot is more robust than either alone.
• Record the singlet frequency as a quality metric alongside viability.
• Not all instruments report width natively; some derive it from area and height, which can behave differently at the extremes.

Frequently asked questions