Table of Contents
- How exactly does percent in gate work?
- An in-depth example to illustrate percent in gate in Cytobank
- Workarounds to get percent of any two populations
- Short takeaway summary
Cytobank offers the percent in gate statistic in the working illustration, gating interface, and export statistics tool. It is useful to understand exactly how the statistic is calculated, especially for the case of percent of grandparent (or great grandparent, etc.).
The first thing to remember is that in Cytobank, a population and a gate are separate entities. Reading more about the difference between a population and a gate will be useful to understand the percent in gate statistic.
To calculate percent in gate, a population and a gate are chosen. The population gives a collection of events according to which gates define the population. From this collection of events, the percentage of those events that fall into the selected gate is calculated. This percentage is the value for percent in gate. No consideration is given to hierarchical relationship between the population and the gate in making this calculation.
Consider the following gating scheme and its equivalent population tree. The gating scheme identifies monocytes that are positive for pSTAT3 in IL10 stimulated PBMCs.
(gating scheme for pSTAT3+ monocytes)
(population tree showing hierarchy to reach the population)
Of the available populations and gates, three different meaningful combinations of percent in gate can be made. Two of these combinations are percent child of direct parent, and one is percent grandchild of grandparent.
- percent of intact cells that are monocytes
- percent of monocytes that are pSTAT3+ monocytes
- percent of intact cells that are pSTAT3+ monocytes (percent of grandparent)
The following image shows the results of each of these three combinations using the percent in gate statistic within Cytobank compared to calculating the statistic by dividing event counts:
(percent in gate statistic compared to percent calculation by dividing event counts)
In Cytobank, comparing parent to child with the percent in gate statistic works fine and gives expected results. However, using percent in gate for percent of grandparent may result in incorrect statistics. The reason for this is the literal nature of percent in gate...A population is chosen, and then the percent of events that fall into that gate is found, regardless of hierarchy. In the example above with intact cells and pSTAT3+ Monocytes, the intact cells population actually leaves non-monocytes in the calculation. Even though the name of the gate indicates monocytes, that gate will still capture non-monocytes if they aren't gated out because a gate stores no information about hierarchy. An illustration of this is below:
(events that would normally be excluded by an additional gate are left in and may fall into the terminal gate)
There are two options for getting the percentage comparison between any two populations.
1) Do the calculation in Excel or other spreadsheet software
- In your Experiment, navigate to the Export Statistics Tool and export event counts for your populations of interest.
- With the resulting file, use Excel or a different spreadsheet program to calculate the percentages of interest.
2) Use the Working Illustration
- (Example configuration in image below)
- Open your Experiment's Working Illustration.
- Use the Population Figure Dimension box to choose the parent population and any child populations of interest to compare to it. Drag and drop the parent population to the first position within the Figure Dimension.
- In the control panel to the left of the Working Illustration, choose a statistic of event count, an equation of fold, and a control such that the parent population is the control.
- Export the Illustration statistics.
(example Working Illustration setup for viewing percentage comparisons between populations to a single control parent)
Use the percent in gate statistic only when comparing direct parent and child relationships. For percent of grandparent or other comparisons that skip generations ("percent population of population"), manually calculate the percentage by exporting event counts with the Export Statistics Tool and doing math downstream in a spreadsheet. Alternatively, use the Working Illustration. See workarounds section above.