Compensation tutorial - FlowJo for Windows

A customer emailed us earlier pointing out that we don't have a compensation process walk-through for the PC version on our website. This entry will hopefully clarify this process a bit. I will be using this data set:

Download Compensation.zip

which is a 3 color Calibur (analog, FCS2) experiment. You should also reference the Compensation Editor manual entry and Mario's compensation 101 page.

The process of compensation should begin with some sort of scatter gate to narrow down the scope of the fluorescence spillover measurement to just the cells of interest:

Tip: Even if you work with homogeneous cell preps a scatter gate is usually a good idea.

In our walk-through, all other gates shown below will be children of the lymphocyte region shown above.

The next step is to create positive and negative gates for all single-stained controls. For the Cy5PE sample, I've created control gates like so:

Tip: don't gate on events outside the dynamic range. That means the gates should not include events on the extreme minimum or the extreme maximum of the histogram.

Tip: if your control antibody doesn't produce a bimodal distribution like above, use two gates that are as far away from each other as possible.

Tip: your control antibody reagent should match the fluor of your experimental probe, and express it as brightly as the brightest experimental sample.

Once you're happy with your gates for the first control sample, do the rest. I've saved my workspace after setting the last gate (right-click, save as)

Download 1.wsp

Tip: There is an alternative approach to setting gates if you are concerned about spillover from one tube to the next. Just graph previously acquired parameter vs. currently acquired parameter - events from previous tube will be high on the axis of the "previous parameter" and you can avoid gating them.

Once all your gates are in place, open the Compensation Editor. You can access this from the Windows menu, or the keyboard shortcut ctrl+F5. This is where you'll agree with me that a scatter gate is a good idea. Notice the interface:

The instructions are to bring each of the neg/pos gates from the workspace to fill out the little white bins. There are two problems with that:

• it is boring
• and error prone

so instead of doing it that way, consider this approach:

1. click on your scatter gate for one of the comp controls
2. click ctrl+comma ("select equivalent gates")
3. drag your selection to the "autoAssign box"
4. notice the white bins are all filled:
   
5. Click "compute and apply". You should get an error message (known bug):
   
   this just means "Hi! You've created a matrix. But it hasn't been applied to any samples yet."
6. to apply the matrix we made in step 5, tab over to "assign tubes". Notice the big empty field? drag FCS files from your workspace here to compensate them:
   
   when I dragged in the sample 3-color from the workspace, the file name gained a little purple bar indicating the sample was compensated:
   
7. Now double-click the compensated file. Notice that the menus for X/Y axes parameter will include the prefix "comp:" - those are the compensated versions of the original parameters.

At the end of step 7 my workspace looked like this:

Download 2.wsp

The plot of compensated data:

and let's not forget display transformation (use the little T buttons alongside your graph window's axes):