How important is proper compensation?

(Go back one page, or back to the Introduction).

If you got some or all of the questions in the quiz wrong, don't feel bad. You will be in the majority of FACS users. The first question that might come to mind is: "why, if so few people know how to properly compensate, have we been getting such good data from flow cytometry for so long?" The simple answer is that most of what we have been using flow to do for the past decade has not required absolutely correct compensation.

For just determining the frequencies of populations, exact compensation is not necessary. This is because, for the most part, the subpopulations that we are interested in enumerating are easily distinguished on the basis of a bright reagent. Thus, a little inaccuracy in compensation, leading to inaccurate placement of the negative population, will not affect the determination of the percentage of cells, since there is still a large valley between the positive and negative population.

For instance, see the figures below. In this hypothetical example, as a continuation from the quiz, PBMC were stained with FITC CD3 and PE CD4 or with FITC CD3 and PE isotype control. The data for each of these two stains are displayed at different compensation values (in pairs, below). You will note that at any compensation value, from none (i.e., completely undercompensated) all the way to somewhat overcompensated, it is still easy to set a gate that will uniquely identify the CD4 T cells (CD3+ CD4+) in order to accurately determine their frequency. However, you will also note that the appropriate gate necessary to quantitate this population is not the quadrant that is set above the unstained cells, but rather it is the gate set on the cells stained with the isotype control and are positive for the FITC stain. This gate is shown in solid lines.

Figure 2: Compensation & cell enumeration

However, proper compensation is absolutely necessary for proper antigen density measurements. Here, any uncorrected spillover will contribute artefactually to the measurement. Proper compensation is also necessary when it is important to distinguish dim populations from negative populations: again, undercompensation will result in overestimating the frequency of the dim cells; overcompensation will result in underestimating the frequency. In general, accurate compensation is also necessary when there are multiple fluors that spill over into a single channel--for subtle reasons beyond the scope of this discussion.

Go on.