Objective: To thoroughly understand the difference between a region and a gate. To know how to construct and apply gates.

In most samples that are run through a flow cytometer there are particles which are not of interest, for example cellular debris, which need to be excluded from the analysis. The way this is normally done is to set a region around a group of visually discrete events of interest in a display, then choose to include or exclude these events using some form of Boolean (AND , OR, NOT) logic. Let's define a couple of terms here,

Regions can be defined as areas drawn on plots displaying flow cytometry data.
Gates can be defined as one or more regions (combined) using logic.

Colour Gating:

Before one starts excluding data from analyses by gating events, it can be extremely informative to visualise the different populations of cells which are encompassed by a particular region by Colour gating. If you haven't already clicked the colour gating example icon above do so now and move the window to the side so you can read this text and look at the graphic. Look at the coloured regions on the top two dotplots, the regions were drawn to enclose the LYMPHOCYTES, MONOCYTES, GRANULOCYTES and EOSINOPHILS in a stained sample of human peripheral blood.

If you consider the scattered light profile and use the information, displayed using the colouring based on the regions, how many population can you elucidate using just the two stains?

Most people with some immunophenotyping experience should be able see discrete CD3+/CD4+ and CD3+/CD8+ stained populations of lymphocytes, that some monocytes bind the CD4 antibody, but not the CD8 antibody, and that some lymphocytes based on the colour of the dot don't bind the CD3 antibody. These are just examples to illustrate the type of information that one can gain by quickly looking at colour gated dotplots. (Note that colours are applied based on their level in the gate hierarchy, see below).
Another example of colour gating can be seen here, this one generated using FCSPress

It is worth reiterating that no events are excluded in these types of analyses, so if the population of interest is infrequent then it may be difficult to visualise in this type of display.

I'd like to now touch on another term which is often used when discussing gating data, that of back gating. Back gating can be defined as the practice of not applying the initial gate to data based on a forward or side light scatter parameters. This approach is relatively common in situations where the scattered light parameters do not show clear delineations of where a population of cells is in the bivariate display space. Therefore one would draw a region on other parameters in the file, normally a "pan" staining antibody such as CD45+ events, then display a dot plot showing the light scatter characteristics of the whole sample, with the leucocytes coloured if they fall within the region (AND logic).

It is however often essential to remove data from the analysis by gating. Gating effectively defines a subset of the data to be displayed. The logic that the software applies, whether it states it up front or not, usual is either AND (inclusive), OR (inclusive) NOT (exclusive). By inclusive it is meant that events must satisfy that criteria, therefore if two regions are draw on a bivariate display, and combined using AND logic, no events can mean both criteria (unless the regions overlap) and so nothing would be plotted. However events could satisfy either region with AND logical gating, and if OR logic was applied to the regions when applied as a gate, then events which are satisfy either criteria would be displayed. This little movie clip illustrates this point.

Let's now talk about Hierarchical gating, and it's implications. Every software package for flow cytometeric analysis which has the ability to colour events which meet a gate criteria, applies some sort of hierarchy. Most packages such CELLQuest (BD-Biosciences) simply read down the list of gates and the event is coloured based on the first gate criteria which is fulfilled. That's fine as far as it goes, however it doesn't address the problem of an event which satisfies multiple gates. One solution to this is a colour eventing approach in which events meeting more than one gate have the colours from each individual gate merged.
Thus start with an image with two colour gates, then apply a third colour gate, and colours get merged . In this example, created with Winlist, a B lymphocyte population which is CD19 positive (coloured blue) is "colour merged" with the whole lymphocyte population (coloured red) and the events which meet both criteria are coloured purple which shows that a good portion of the B cells have a lower FSC than the majority of the lymphocytes.

To be continued when time permits....

Click here to go back to Histogram discussion:

or be a "glutton for punishment" and

Click here to go back to Regions and Gates.

FACS HOME