Attention: Confluence is not suitable for the storage of highly confidential data. Please ensure that any data classified as Highly Protected is stored using a more secure platform.
If you have any questions, please refer to the University's data classification guide or contact ict.askcyber@sydney.edu.au
Guide for spectral unmixing in Flow-Jo
- Edwin Lau
- Suat Dervish
Why use spectral unmixing?
Spectral compensation utilises signals from non-primary detectors. The additional information allows the software to separate individual colours more precisely, decreasing spillover.
The more light (data) you collect, the better the unmixing can perform. It can be done on any cytometer as long as you collect n+1 data points where n is the number of colours in your cytometry panel.
Recording single stained controls:
You’ll record single stained controls just like conventional compensation. The same rules apply:
Positive controls have to be equally bright or brighter than the sample
All signals are within the linear detection range
The positive and negative populations must have the same auto-fluorescence (i.e. Pairing beads with beads with the same lot number, cells with cells)
Fluorochromes used for compensation must be the same as the sample (Can’t compensate AF647 with APC for example)
Additionally you’ll want to:
Keep the parameters you are not using. Do not delete them!
Check if signals in their primary detectors are the brightest. If signals are brighter in non-primary detectors we have observed non-optimal spreading occurring.
Spectral compensation in FlowJo
To use the spectral compensation feature, you’ll need FlowJo version 10.6 or later (FCS express now has this feature also)
Drag in the single stained controls.
Click the compensation icon and select the parameters you are using. Flow Jo should pick the positive & negative populations automatically. Adjust if needed.
Check the spectral box. Click on selected detectors select all fluorescent parameters.
Click on view matrix for calculation. You can then drag and drop the [M] icon to apply spectral compensation onto you sample.
Optimise weights may help to reduce spread but it’ll take quite some time for calculation, depending on your compute power.
What if spectral compensation looks worse than traditional compensation?
This can be due to sub-optimal PMT voltage settings.
Voltages set by CS&T are sometimes too high, especially on high-parameter instruments. High secondary PMT voltages can result in increased population spreading and therefore reduced signal separation, even if that detector is unused.
If you observe increased population spreading following spectral unmixing, you probably need to check for high PMT voltage on unused detectors. A good indication is when the non-primary detector signal intensity is higher than the primary.
Under-performing detectors (high background noise, low linearity for example) may also throw off the matrix. Identifying and leaving that detector out can help.
Reference video: