From my experience, there are a couple of reasons why your experimental data (slice4_gated) might look uncompensated even after editing the spillover matrix based on compensation controls. Here's why:

1. Live/Dead Fixable::APC Spillover:

Your current compensation strategy only addresses spillover between FITC and PE. However, you're using a three-color panel with Live/Dead Fixable::APC. This fluorochrome (APC) can potentially spillover into both FITC and PE channels.

• Solution: You might need to acquire single stain controls for all three fluorochromes (FITC, PE, and APC) to create a complete spillover matrix. This will allow you to compensate for APC bleed into both FITC and PE channels.

2. Insufficient Compensation Adjustment:

While you adjusted the coefficients to get zero PE signal in the single-stained FITC control, it might not be enough for your specific sample. Remember, compensation beads are designed to be more uniform in staining than actual cells.

• Solution: You can try further refining the compensation matrix. Here are some approaches: Use a more advanced compensation method like fluorescence minus one (FMO) controls. FMO controls involve staining cells with all antibodies except one, allowing you to estimate the spectral overlap from other fluorochromes. Utilize software tools that offer automated compensation based on your single stain controls. Consult your flow cytometry core facility or instrument manual for specific recommendations on compensation strategies.

3. Population Shifts:

The diagonal distribution in your FITC-PE plot for the experimental data could indicate a population shift compared to your compensation controls. This can happen due to:

* Different cell types in the sample compared to compensation controls. * Variations in antibody binding efficiency between controls and your tumor sample.

• Solution: Unfortunately, compensation cannot entirely correct for population shifts. You might need to consider alternative gating strategies to isolate specific cell populations of interest before applying compensation.

Here are some additional tips for flow cytometry compensation:

• Always use fresh compensation controls for each experiment.
• Ensure your instrument settings (laser power, voltage settings) are consistent between compensation controls and your actual sample acquisition.
• Validate your compensation using positive controls (known positive populations) to ensure accurate separation of populations.

By addressing these points and potentially refining your compensation strategy, you should be able to achieve a more accurate representation of your experimental data.

Hey Elena! Idowu Goodnews Oluwaseyi already gave some nice tips about things you could adjust in your experiment, maybe I can add a little more to it.

Firstly, I highly recommend to measure a single staining for your live/dead marker - in general you should always prepare single stainings for all channels / fluorochromes used. Since most of the beads used for single stainings do not bind fixable viability dye (there are a few options now, but I have not tested this so far), we use cells for that and of course an unstained sample of cells. Given that you analyze your data with FlowJo you can specifically select the unstained cells as negative control for (only) the live/dead stained single staining in the compensation setup window. That should help further to obtain good values in the compensation matrix.

Secondly, you could optimize the fluorochromes you are using for your panel. FITC has some bigger spillover into PE, so a different combination of fluorochromes could do the trick. Or even better: Your flow cytometer is equipped with a 561nm laser. Then you could measure FITC in a detection channel of the 488nm laser and PE in the detection channel of the 561nm laser (yellow-green), where FITC excitation (and therefore spillover) is pretty much zero. Maybe you could let us know on what flow cytometer you are acquiring your samples.

Lastly, have you done any pre-gating on your sample "slice4_gated"? To me the diagonally distributed cells seem mostly to be those that are also positive for APC (i.e. your dead cells). I wonder whether a pre-gating on live cells would improve the outcome in the bivariate FITC-PE plot.

I hope that helps.

Henning Düsedau thank you very much for your reply! So I did also measure a controls for live/dead with APC fluorochrome, both stained and unstained. The matrix that I show here already includes compensation for APC, I mostly focussed on FITC/PE spillover since its where I see mor inconsistencies (I edited the matrix but nevertheless).

For now I was using a BD Canto II with 488 and 633nm lazers (would be optimal for me since its in our lab), but I will probably switch to Fortessa in core facility to have more color options.

I will try excluding dead cells to analyse FITC-PE :) Again thank you very much!

Adding to all the previous suggestions:

When staining cells to generate an additional "compensation control" for your live-dead stain, you will want to have a clearly dead cell population that will be definitely stained. Otherwise, healthy cells have a high chance of only being weakly stained, especially if the reagent reacts with the dead cells only.

So you may want to take half of your sample and apply 60°C heat to it for 5-10 minutes, then spin it down, resuspend in fresh medium and mix 1:1 with live cells. Make sure that your cooked cells are cooled down before mixing it with live cells. I got this advice from a FACS core facility expert and have found that it solved a lot of my compensation troubles.

What she also recommended is doing the above procedure (mixing dead and live cells) and titrating your live-dead reagent. Most of the times manufacturer recommended amounts of reagent are extremely high and almost excessive, leaving you with poor live-dead population separation and high spillover in other channels. So you might want to see if using less live-dead reagent is possible, will also be great budget-wise.

Hope this helps!

Sofya Novikova thank you for your reply! I did almost the same, however I resuspended my both live and (to be) dead cells in PBS and left the live ones on ice while the others were killed with heat. I also titrated the Fixable Live/Dead staining (1:100, 1:500, 1:1000), however I dont really want to go lower than 1:1000 because I'm really uncomfortable pipetting less than 0,5 µL volume :) Again thank you!