Q: What are display transformations?

A: Display transformations change the traditional log scale to allow it to have a zero and a negative region. This is accomplished by converting a small region of the scale at the axis to linear. The display transformation allows you to display populations of cells that are squished against the axis as nice round clusters.

Q: Why should I use display transformations?

A: When data is properly compensated, it is common that a large number of cells are displayed squished against the axis. The cells become piled up in the first channel (against the axis) because the fluorescence parameters are displayed on a log scale where it is not possible to display "zero" or negative values. Click here for more details about why this occurs. FlowJo's display transformation displays your data on an altered scale that has a zero and a negative region.

Digital Data can also be transformed to smooth out the picket fences that occur in the lower regions of the log scale. Click here for more information.

Q: What does it mean to have "less than zero" fluorescence?

A: This is not meaningful. There is no information to be gained by comparing a population with values less than zero to a population of cells with fluorescence values in the first decade. These are in the same population.

Q: Is this altering my data? Is it legal? Is it ethical?

A: The data is exactly the same before and after transformation. The fluorescent values for each cell are not different in the two different displays. Therefore, the statistics you will calculate on the data are the same as well. The display transform simply allows you to view the negative populations as nice symmetrical clusters instead of squished against the axis. There is no illegal or unethical altering of your data.

Q: The double negative population comes out looking squished if you transform the data. The single positives look nice, but there's often a sharp edge to the population in the double negatives. Why?

A: First of all, make sure that you calculate the transformation on a gated population (like lymphocyte gated, or whatever represents the major population you are looking at). The presence of other cells (especially, dead cells, if you are including PI) will significantly change the appearance of the transform in ways that are not desirable, since it is trying to accommodate all of the cell populations, even those you are not interested in viewing.

Second, the sharp edge of the population is a consequence of using data that has all "negative" events pinned to the left axis. If you are using BD's digital software (i.e., DiVa or LSRII), then you should choose to "Export Experiment", which does not pin all the events on the axis--you will get smoother distributions. Otherwise, you will see this effect because of the large "plug" of events that are all on the axis before compensation (compare the uncompensated parameter versus the compensated parameter). The events that started life on the baseline stay with that imprint even after transformation, while the events pushed onto the baseline by compensation can be visualized below it after transformation.

Q: Is the solution to the "edge" phenomenon to increase the voltage a the time of acquisition so that the negative events are sufficiently far from the baseline to look "rounded" after transformation?

A: Yes, as long as you don't push your positives off-scale.

Q: The gates on templates created in previous versions (before the transformation feature was implemented) do not reflect the transformation.

A: Since this problem isn't a simple one and it only arises once: during the transition from not having the transformation ("old workspaces") to now having it ("old workspaces where you specifically now tell it to transform"). Since it is an active process to transform old analyses, you need to check and adjust your gates.

Q: Is the new version of FlowJo backcompatible - that is: can workspaces set up and compensated with version 4.2 and earlier have the transformations defined in 4.3?

A: Yes.