Autofluorescence in Flow Cytometry: Why It Happens and How to Reduce It

Your unstained tube is sitting two decades up the FITC axis, your “negative” population has no clean edge, and a panel that worked on lymphocytes falls apart the moment you run macrophages. That background glow is autofluorescence — signal from the cell itself, mostly from flavins and NAD(P)H, not from any antibody you added. It peaks in the blue-green range where FITC, GFP, and PE live, and it falls off sharply past about 600 nm. Most autofluorescence problems are not bad luck; they are a handful of avoidable mistakes. Here are the ones that cost the most time.

Mistake 1 — No unstained control matched to the sample type

The single most common error is reading a stained sample without an unstained tube of the same cell preparation. Autofluorescence is cell-type-specific: granulocytes, monocytes, and macrophages glow far more than lymphocytes, and tissue-digested or plant samples more still. A lymphocyte unstained control will not tell you where the negative sits for myeloid cells in the same tube.

Fix: run an unstained control for every distinct sample type, not one for the whole experiment. It defines where “negative” actually is and lets you judge whether a dim positive is real or just baseline glow. This is the same logic behind using FMO controls for borderline markers — the control sets the boundary, your eye does not.

Mistake 2 — Blaming compensation for an autofluorescence problem

When a negative population smears upward, the reflex is to add more compensation. But autofluorescence is not spillover. Spillover is one fluorochrome leaking into another detector and scales with that fluorochrome’s brightness; autofluorescence is intrinsic cellular emission that appears even with no fluorochrome present. Compensating against it just drags your real negatives below zero and corrupts everything downstream.

Fix: diagnose first. If the unstained control is already high in the channel, the problem is autofluorescence, not the matrix — turn to sample prep and fluorochrome choice. If the unstained control is clean and only the multicolor sample smears, that is a spillover or spreading question, covered in over- and under-compensation.

Mistake 3 — Putting your dimmest marker in the autofluorescence channel

If you assign a rare or dim antigen to a FITC-equivalent green channel on an autofluorescent cell type, the marker drowns in background before you start. Brightness matching exists partly to avoid this.

Fix: push dim antigens to brighter fluorochromes and, where possible, to channels above 600 nm where autofluorescence is low — APC, the red-excited tandems. Reserve the green channels for high-density antigens that can outshine the glow. Checking a fluorochrome’s emission against the autofluorescence-heavy region before you order antibodies is exactly what our fluorophore spectrum viewer is for; it flags green/yellow-channel choices on autofluorescent sample types. The brightness-vs-density logic ties directly into multicolor panel design.

Mistake 4 — Leaving dead cells in the analysis

Dead and dying cells bind antibody nonspecifically and autofluoresce strongly across the spectrum. A few percent dead cells can fabricate an entire fake positive population.

Fix: include a viability dye and gate dead cells out early, before any marker interpretation. Removing debris and dead events is the cheapest autofluorescence reduction available, and the dye choice matters depending on whether you fix — see live/dead discrimination with DAPI, PI, and fixable dyes.

Mistake 5 — Ignoring what serum and fixation add

Fetal calf serum absorbs and contributes background in the violet-blue range, and aldehyde fixation raises autofluorescence in its own right — a fixed sample is not optically identical to the live one you titrated on.

Fix: standardize FCS concentration in your staining buffer, or switch to BSA if the background persists; keep fixation minimal and consistent across the experiment so the baseline does not move between tubes.

The spectral option

On a spectral instrument you do not have to fight autofluorescence purely at the bench. Because the full emission spectrum is captured, autofluorescence can be treated as its own “fluorochrome” — an unstained or autofluorescence reference is included in the unmixing, and its signature is extracted out of your markers. It is the most effective handling available, but it depends on a good autofluorescence reference and is its own source of error when done carelessly; the failure patterns are in spectral unmixing gone wrong.

Spot-check yourself

Before you trust a dim positive, run the list: is there an unstained control for this sample type, are dead cells gated out, is the dim marker on a channel where autofluorescence is low, and is the unstained tube — not the compensation matrix — what is driving the smear? Most false positives in autofluorescent samples fail at least one of these checks. The historical primary reference on the problem, Mosiman et al., Cytometry (1997), is still worth reading for why the green channels are the hard ones.