Hello,
I did some flow cytometry acquisition a week ago. Now, I finally got the chance to analyze the data and I unfortunately discovered that one of my compensation controls didn't work well: not enough events were acquired, PMT voltage was set too high, and positive events (dead cells) don't appear (see attached pictures). Therefore, my compensation matrix doesn't make any sense (pictures attached).
I was thinking if it is okay to apply a compensation matrix that is the result of an old acquisition (two months) of compensation controls of the same panel, that worked perfectly? I know for sure that it is not the best thing to do, but giving the circumstances is it a tolerable thing to do? I mean I will for sure have to repeat the experiment and re-do the acquisition, paying more attention to the quality of my compensation controls, but I was hoping to make some use of this data.
Thank you,
Martin
Hey Martin, it's tough to say what's going on here, but I'm guessing that the shoulder on the left in your first histogram corresponds to the FSC low SSC low events. It's not ideal, but does your compensation matrix become useable if you call the main peak to the right your positives and the shoulder on the left your negatives? If you expand your first gate to include all events except for the debris, does that give you enough events with which to compensate? I would try tweaking the current matrix rather than using an old matrix that was probably not acquired with all of the same voltages.
Depending on what remained the same in this experiment, vs the old, it can be appropriate to use the old compensation matrix. If auto fluorescence (different cell line, different intervention for apoptosis), and/or PMT voltage was lowered, then the old compensation matrix will not work well.
Im also curious about the two shoulders. In the FSC and SSC gating plot it looks like there are two cell populations; low SSC/FSC, and low SSC/high FSC. A lot of interventions that cause apoptosis cause the cell size to increase, which would correspond to the low SSC/high FSC pop. Further, the low SSC/high FSC pop has a MFI greater than 5x10^3, which is typically where i start to see postive cell populations (although this depends on the flow cytometer). So its possible you have 20% of cells that did not enter apop, and 80% that did, or incubation time with your dye was too short.
Hence follow up question; how did you induce apoptosis, and did this change compared to the previous time? Is it a new lot of AmCyan? How long were the cells incubated in AmCyan for this experiment, compared to the previous? It may be AmCyan incubation time was not sufficient this time to stain all cells.
Further, too high PMT voltage is not as bad as too low. As long as you are in the linear scale of your detector. Given that your signal is 10^3-10^4 MFI you are almost certainly in the linear scale. Although not ideal, I wouldn’t worry too much about the voltage being too high in this scenario.
Gion Fränkl I did not induce apoptosis. Nothing changed compared to the previous experience, although I have to check to see if I did use the same dye lot. Besides this, I have the same cell population, same incubation time etc. The incubation time is 20 minutes at RT as recommended by manufacturer. The problem here (at least for me), is that I dont have any (or very few) positive events (dead cell aka AmCyan stained cells). Therefore, I can't calculate an accurate compensation matrix.
@Martin Schnitzler I misunderstood you then, I incorrectly assumed by positive events you meant postive cell death sample. In this case i guess a vast majority of your cells are alive, which is not great for generating the compensation matrix as you said.
With the info you gave me, and considering the AmCyan is a small molecule (should be really no difference between lots, unlike antibodies), you should be able to use the same compensation matrix. Once i have a defined cell population, defined panel, and set PMT voltages, I use the same compensation matrix for different patients. You have changed none of the important variables that could make a compnesation matrix inapplicable (excluding PMT being higher, which is not a big issue, especially if all your cells are alive, which means little to no AmCyan to excite and thus have emission spillover to other channels/photodetectors).
As far as i know using a compnesation matrix from one day on experiments on following days is standard, unless you change experimental parameters. Have you been historically making a new compensation matrix for every experiment, even if parameters dont change?
Ideally you would repeat, as you already mentioned. But for analysing this experiments data with an older compensation matrix should be fine
- Badges
- Science method