Dear Su,

Generally, during the staining experiments with FACS the gating plays important role. I hope you are using log scale. First set the gating woth respect to negative and unstained control. According to gating, the peak shift with fluorescence will show the population positive for staining after stress/treatment.

As you are saying that there was an obvious peak shift with fluorescence, someone can guess that after particular treatment of yours, cells may be altered in their complexity (scattering). So, I suggest you to check under microscope for their shape/morphology/clusters/..etc to get an idea about the obvious peak shift.

Good luck..!!!

Dear Golla,

Thanks for your response. As you said gating is important, but what after I get the positive rate of both control and treated cells, for example, 35% and 42%, does that mean the difference are significant?

And also thanks for your advice for peak shift in flourescence. Actually the dye is an indicator of ROS, the shift represents an increase in total ROS after my treatment. What I want to figure out is whether this shift is "big" enough to say there is significant increase of ROS caused by my treatment.

Thanks again for your response.

Dear Su,

That significance depends on the reproducibility and the shift difference. If it is from 35% to 42% surely one will not consider significant shift. If the shift is more than 20-and above % of population shift might be significant if reproducible. Apply suitable significant tests to confirm your hypothesis.

Good luck..!!!

Dear Su,

besides applying gating, you can also try to compare the median of the staining intensity. The results that you get from gating stronlgy depends on the gates you have chosen. You can get a shift from 38-42%, which does not sound too impressive or with a higher value for the boundary you might get numbers increasing from 5 to 20 % or so (that sounds more impressive). The medium staining intensity cannot be manipulated in this way. With these numbers (% cells from the gating or the median) you should run the usual tests for statistical significance (ANOVA). But this is only a number. I guess it is more important to know if the changes are indeed biologically significant. To assess this, we usually compare the staining with the result that is acchieved by imposing exogenous oxidative stress. This can be simply done by adding hydrogen peroxide. Usually 1 mM will have a significant effect. In this context it would also be interesting to know if the treatment kills the cells, which can be analysed in parallel with PI staining.

But what about your 96well assay? How much was the increase in fluorescence with this technique - how does this relate to the flow cytometry results?

Best wishes,

Norbert

Dear Rob,

Thanks for your explanation. In my situation, the median, mean and geographic mean for control was 58, 71, 56; while for the treated group, there were 66, 76 and 58, with the positive control (H2O2 treatment) 83, 128 and 88. I will surely repeat my experiments, but I also would like know what you think on the data I have got.

Dear Norbert,

I agree with you that I should add a positive control by exposing cells to H2O2. Actually, my 96-well plate result was more or less the same like flow cytometry, but the increase in the treated group was a bit more "significant" than the flowcytometry. I guess I would have to wash more times next time I do the FCS, because significant background could be observed under the microscope in a separate DCF experiment when I just washed cells for once. three more washes with pre-warmed PBS could significantly reduce background.

Hi Siyuan. Relying on % counts when using fluorescent probes is indeed a challenge. It is usually difficult to determine the positive-negative boundaries. Te best way should be to assess the median fluorescence intensity (or the fold change of it). If your biological replicates slightly shift from basal or control conditions you still would be able to find significant differences using standard tests for non-gaussian distributed populations (remember that flow cytometry histograms do not usually exhibit a gaussian behavior) no matter how small that shift is as long as you have enough biological replicates.