Question 12 of 20 Intermediate Concept
How does spectral flow cytometry differ from conventional flow cytometry?
Short answer
A conventional cytometer gives each fluorochrome one detector behind a filter. A spectral cytometer records each cell's full emission across many detectors and every laser, then unmixes the signal using reference spectra, which lets it separate dyes with similar peaks and subtract autofluorescence.
Why
In conventional instruments the rule is one detector, one marker. Overlap into other detectors is removed by compensation. A spectral instrument spreads the emitted light with a prism or grating over a detector array.
Each fluorochrome then has a signature across all detectors, not just a peak. Software fits every cell's measured spectrum as a mix of the reference spectra, usually by least squares. Ferrer-Font and colleagues point out that dyes with the same emission maximum but different off-peak shapes can share a panel, which allows larger panels.
Autofluorescence can be measured as its own signature and removed. That helps with tissues such as lung, where macrophages are highly autofluorescent. Unmixing depends entirely on the reference controls.
Each must use the exact fluorochrome, and for tandems ideally the same lot, as the sample, and must be at least as bright. Unmixing does not remove spreading error. Two highly similar dyes on co-expressed markers still degrade each other's resolution.
For example, a 30-color panel on a spectral analyzer still needs careful marker assignment and single-stained controls for every dye.
A strong answer covers
- Full spectrum across all detectors and lasers
- Unmixing against reference spectra replaces compensation
- Similar-peak dyes with different signatures can coexist
- Autofluorescence can be extracted as its own signature
- Reference controls must exactly match the sample dyes
Common mistakes
- Assuming spectral unmixing removes spreading error between similar dyes
- Using a spectrally similar substitute dye as the reference control for a tandem
- Believing panel design rules no longer apply on a spectral instrument
The interviewer might ask next
How would you check whether unmixing errors are creating a false population in a spectral dataset?