Compensation, spreading and FMO: three different questions

• Explain why correct compensation does not collapse every negative distribution.
• Distinguish a single-stained compensation control from an FMO control.
• Recognise when panel context undermines the interpretation of a dim marker.
• Describe an appropriate conclusion when full-stain and FMO distributions overlap.

Editorially approved by Juan Manuel Ojeda. Independent scientific review has not been completed. Compensation corrects fluorescence spillover: a fluorochrome can contribute signal to more than one detector. Measurement variability remains after that mathematical correction. A widened negative population is therefore not, on its own, evidence that compensation is wrong [1,2].

Single-stained controls provide information for estimating spillover. An FMO contains the panel's other reagents but omits the reagent being evaluated. It helps show the background distribution in the multicolour context; it is not a substitute for the controls used to calculate compensation [1,3]. Because the omitted antibody is absent, an FMO also cannot directly measure nonspecific binding attributable to that antibody.

Spreading is an interaction between the measured signals, optical configuration and panel context. A spillover spreading matrix provides information different from the spillover coefficients themselves [2]. A dye assignment that performs well for a bright marker can still make a dim co-expressed marker difficult to resolve.

CFCM working question: do these data support separation of the proposed positive population from the relevant negative distribution? That question is more useful than asking whether a cloud fits neatly into an aesthetically pleasing quadrant.

Spillover: signal from one fluorochrome measured in another detector. Compensation: mathematical correction using an appropriate model and controls. Spreading: remaining distribution width associated with measurement uncertainty. FMO: a contextual negative control, not proof that every event above a chosen line truly expresses the marker.

Synthetic panel-design scenario. A full panel appears to contain a small CD25-dim extension. It sits within cells strongly labelled by another reagent. The corresponding CD25 FMO shows a similarly shaped extension in that same parent population. The single-stained controls do not indicate a residual spillover trend requiring a correction.

A reasonable next step is to compare the distributions within the same biological parent, using identical acquisition and display settings. Consider whether the panel resolves the question, rather than moving the CD25 boundary until the desired frequency appears. A biological control known to provide a suitable contrast, review of titration and an alternative dye assignment can help discriminate competing explanations. These are investigation options, not an automatic instruction to change a validated panel.

The observation does not prove that every candidate CD25-positive event is an artefact, nor does it establish absence of CD25. It indicates that the current evidence does not demonstrate a clearly resolved population in this constructed example.

• Adjusting compensation merely to make a negative cloud narrow.
• Using the unstained tube as though it contains all multicolour background effects.
• Estimating a compensation coefficient from an FMO rather than an appropriate single-stained control.
• Using different parent gates, transforms or detector settings for the comparison.
• Calling overlapping distributions evidence of either biological absence or definite expression.

During panel development, record the biological question, expected co-expression, candidate fluorochromes, control strategy and the evidence that the critical distinction remains resolvable in the full panel. Any clinical implementation or alteration still requires the laboratory's change-control and validation process.

A colleague says: “The FMO has events in the positive quadrant, so subtract its percentage from the full-stain percentage and report the difference.” Is that automatically justified?

1. Can correctly compensated data show a broad negative distribution? Yes.

2. Does an FMO directly test the omitted antibody's nonspecific binding? No; that antibody is absent.

3. Does the synthetic example prove CD25 is absent? No. It questions whether the current panel resolves the proposed distinction.

[1] Roederer M. Spectral compensation for flow cytometry: visualization artifacts, limitations, and caveats. Cytometry. 2001;45:194–205. Original technical paper.

[2] Nguyen R et al. Quantifying spillover spreading for comparing instrument performance and aiding in multicolor panel design. Cytometry A. 2013;83:306–315. Original study.

[3] Cossarizza A et al. Guidelines for the use of flow cytometry and cell sorting in immunological studies (third edition). Eur J Immunol. 2021;51:2708–3145. Open-access guidelines. The guidelines distinguish FMO boundary/spreading information from compensation controls and explicitly note that an FMO cannot account for nonspecific binding of the omitted antibody.

The CD25 scenario and proposed investigation are CFCM teaching constructions, not a patient case or published experimental result.