The Graph Window, Part 1

aspect ratio pixel

The fun stuff about FlowJo is just getting in there and opening a graph window, right? The graph window facilitates visualization and graphing. It shows two measured parameters by default in a pseudo-color dot plot display. The parameter label's here, which says Comp-Ax700-A CD3. That's the CD3 distribution. Here's Comp-APCH7-A HLADR. That's the HLADR distribution. I'm just visualizing two different measured parameters here and looking at the distribution of the events in that file.

The graph plot initially shows a 2-D representation of the data. There are gating tools up in the top left-hand corner that you can use to draw a geometrical gate around a population of interest, defining that as a certain marker positive. For example, the CD3 positive population are my T cells. There are also plot view options down below and active gate options so you can change the way this gate is displayed and the way the data is displayed. I'll show you that in just a moment.

When I change the type of data display, I can do black and white plots like contours, densities, zebras, and dot plots, or even one-dimensional histogram profiles. I usually like to use pseudo-color for gating because it's nice and pretty, but if you're going to have page charges in the publication you might consider using a dot plot or a contour plot instead. I'll show you how to change those in just a second.

There are many different gate tools. There are ellipses, an automatic gate that follows the contour line of a population, or my favorite is a polygon where you just click. Every time you click the button you get a point, and you can make any number of sides of that polygon. Any shape that you want. Remember that there's an undo button, so if you don't like what you just created, you can just click on the undo. It'll step you back. Then there are navigation buttons up in the top right-hand corner that'll move you up or down through the gating tree or left and right will move you through the samples in the sample list in the workspace.

Let me go ahead and draw some gates with you guys. Again, feel free to ask any questions about the details of gate creation or the graph window as I go through it. I'm going to start with this sample LD1 no stim, no stim. That is just the background control stained with all of my reagents in this panel. If I double click on the sample name, it opens up a graph window displaying those events. Here is my graph window. We're looking at forward scatter on the x-axis versus side scatter on the y. Forward scatter is a relative measurement of cell size. Side scatter is granularity or interim membrane content. These are my major population that shows the red dots is my lymphocytes population. That's what I'm going to want to isolate.

The pseudo-color nature of this plot display shows up many events over the density events at a particular position within the graph is color coded. Red means there's lots of events at that position. Then yellow, then green, then light blue, and dark blue to the outliers. I'm going to make a couple gates here with you guys using these gating tools in the top left-hand corner. First thing I'm going to do is clean up this sample a little bit with a gate to discriminate between double cells that may have passed by the detector two at a time and single cells.

You can do this when you collect forward scatter area versus forward scatter height on your instrument together. Then display both of those parameters. There's an area scaling factor set on the machine that allows you to see all of the single cells on the diagonal. Then, anything that's off the diagonal is going to be considered a double cell. It might be two different cells, different types, and I don't want to keep those cells. What I'm going to do is use my polygon tool and create a gate. When I double click, it closes the polygon and brings up a naming field where I can then type in a name for this gated population that I'm creating here.

I'm going to call this one a "Singlets Gate". S-I-N-G-L-E-T-S. When I press return, it shows me the actual graph. I'm sorry, the gated population name, the gate, and the statistic here, which is the frequency of parent statistic. That's the basic statistic that you get when you create a gate in FlowJo. Frequency of parent means that there's the percentage of cells or events contained within this child gate from the parent population. You'll notice that when I create that gate, it also shows this gate as an indented population node on the sample within the sample's pane of the workspace.

In order to move down into that population of interest that I've gated, I'm going to double click on that gate that I created, and it opens up a new graph window isolating those events. I can also do that from the gated population in the workspace. Double click on the population, brings up a graph window containing only those events.

Now I'm going to make another gate on a forward scatter versus side scatter using my ellipse tool here. Just click on the tool in the top left hand corner, and in this case I'm clicking once, dragging with the mouse button held down, and then when I release it creates the gate and brings up the name field. I'm going to call this a "Lymphocytes Gate". I'm just isolating the small, forward side scatter cells that are my major lymphocyte population and excluding all of the stuff that's larger that would include any of my monocytes and other larger cells. 

There's my gated population. Remember that you can always change the population, the gate that you created, by grabbing the handles on either end and moving the sides. You can move it around, and you can that the statistic there automatically calculates. You can always modify the gate after you've created it and positioned it just the way you want it. Here I'm going to show you that we can also display different types of graph plots under the options dropdown menu. You get a type. This is a type menu that says pseudo-color, but I can make a contour plot. If you do that, be sure to check the "Show Outliers" to get the outliers, make it look pretty. There's also density plots, zebra plots, monochromatic dot plots, which down sample to the number of events you put into the box here, and then one-dimensional histograms.

If you do a histogram profile, you'll notice that the tools change for gating. There's just a range gate tool to define a positive or negative population or some region of that histogram, and then a bifurcation tool which sets the center and creates two linked gates, a positive and a negative, on that distribution. If you don't like what you've done, you use the undo button. That'll get rid of them.

I'm going to go back to pseudo-color here, and I'm going to answer that question. Somebody raised their hand. How do you come to the main gate? I'm not quite sure there. The easiest thing to do, to get started David, is to double click on a sample. That will show you this marker profile, and it should show you forward versus side scatter. To create the gate, you click on the tool up in the top left-hand corner such as this ellipse tool. Then, go to the graph window, click in the graph window, and then drag across to the other side of the population that will open up this ellipse. Release the mouse button, and it'll create the gate giving you that naming field. Okay?

Then if you want to move into that population, you can double click on the gate, and that will open up a new graph window. There we go. We'll open up a new graph window containing only those events, and then you can look at a different distribution and make your second gate. By doing that in a series, you build up a gating hierarchy or an ancestry that populates or that parses down your samples into distinct populations based on multiple distributions of different markers.

Let's go ahead and do that a couple times here, and I'll try and move quick. I'll show you active gate options in just a second. Well, let's go ahead and do it. Active gate options. Another menu here. Notice how they're grayed out if I don't have anything selected? This moment I select a gate by just highlighting it clicking it once, it shows me these options, and I can tint the gate, I can change the color of the gate here. I can make it larger or smaller with the plus and minus buttons. You can choose to gate events inside. That's the default of every gate, but you can also uncheck that it and it toggles on and off. Now, without the events inside checked, I'm actually excluded all of the events within that gate. See, the number changes there.

Then I can make this gate magnetic. If you have lots of drift of your samples, and you just want this gate to center on that major population, the magnetic option will take that gate and it'll drag it to the densest portion of the population and show me a factor of how far that moved. I usually prefer to leave that off and just center things in a static way. Those are some different active gate options that you can use to change and modify the gating.

Let's go ahead to this other population. Now I'm on the lymphocytes here, and I'm going to go ahead and show HLADR versus CD3, with CD3 on the x-axis. You can see this distribution. In this case, I'm going to use a polygon tool, and I'm going to say, "I want to cut through this distribution right here and define these cells over on the right as my CD3 positive." If I double click, it gives me a name field, and I'll say the "CD3 Positive" cells.

This distribution looks really good, but if it doesn't, there is a button called the T, Transform, button next to either of the axis label menus. Here's the axis label menu where you can change what distribution is being displayed. Here is the T button. If you go to T and use the "customize axis" option, it opens up a one-dimensional histogram of a 2D plot of the parameter that I initiated the T button from. I can change the scaling of that parameter display. What we usually have is what's called a bi-exponential transform function, which means you have a linear range around zero and log further out. You can change how much space is on the top end or the bottom end of that distribution to make it get it on scale. Then you can change how much compression around zero occurs. If I go to the right with my width base slider, it compresses the data around zero. If I go to the left, it spreads it out. Using those options, I can make this data look very bad.

Actually, what I'm going to do is apply this change not just to the CD3 Ax700, but I'll apply it to the APCH7 distribution in the y-axis too. When I apply those changes I've made by removing all the extra negative space and spreading the data out too much, then you can see it looks like the data's squished on the axis and I can't see everything. If I go the other way and say, "Make lots of extra space," in the back end and compress the data too much, well that doesn't look very good either. I wouldn't even be able to accurately draw a gate between the positive and negative if I was looking at it that way.

I would encourage everybody it explore the T button customize axis options where you can use the slider bars and the plus and minus buttons here to scale your data set just the way you like and make it really pretty. When I apply those changes, then I can see all the data here. You'll notice that the gate automatically scales with the transform process. You can find that perfect distribution where you can accurately gate between positive and negative using the T, transform, button customize axis options.