[Cytometry] CD150 negative population shifts down! How is this possible?
roederer at drmr.com
Sun Jun 21 16:01:02 EDT 2009
I haven't seen your attached picture (it wasn't attached).... but I
have a pretty good idea of what is going on. We first noted this
behavior some years ago, in the very early days of BD's digital
electronics, and it hasn't gone away. We've come to affectionately
refer to it as the "baseline restore issue"... While we are not
positive that we understand what's going on, we have explored this
issue a great deal, discussed it with BD and others, and think we have
an explanation. And we have at least 2 partial solutions:
(1) Use additional washing steps of the cells after staining; and
(2) Ignore the problem
We originally saw this issue with QD655 staining (which is very
bright, like PE). However, we also began to see it occasionally with
other reagents, but usually only those that are very bright. We
explored this issue quite a bit with a CD8 reagent (which is bright on
cells). We found, first of all, that it didn't happen with all
samples -- it seemed to be sample specific. (More on this later).
Second, the extent to which the negatives crept down (towards more
negative values) depended on the brightness of the reagent (use lower
concentrations to lower the positive values, and the effect
diminishes) as well as the frequency of the brightly-positive events
(mix unstained cells with stained cells to varying extents, and the
effect diminishes as well).
Ultimately, I believe that what is going on is that the "baseline
restore" on the instrument will sometimes over-estimate the background
fluorescence. Baseline restore is one process by which the instrument
calculate the background of the stream as cells go through, in order
to determine the actual fluorescence of the cell when it goes
through. Thus, as the background in the stream increases, the
instrument subtracts more from the cell's fluorescence in order to get
the actual cell-associated fluorescence.
The background is estimated as the average fluorescence outside of any
trigger events, over some recent time frame before the cell entered
The problem with this approach is that if you have very bright clumps
of fluorescence that go through the system without a trigger happening
(because they don't have enough scatter signal, i.e., they are
debris), this can artefactually increase the baseline dramatically.
The next cell that comes through will have WAY too much fluorescence
subtracted from its measurement, and will go way negative.
This hypothesis is supported by our observations: diluting the sample
with unstained cells will dilute the debris significantly, lowering
the rate at which this happens. Using less-brightly-staining reagents
will to a large extent lower the fluorescence of any such debris/
clumps, also diminishing the effect. And finally, the problem tends
to occur more with "dirty" reagents, for example, our early home-
conjugated Quantum Dot reagents that have a lot of aggregates.
We hypothesized that if this was an issue associated with debris, then
additional cell-washing steps following staining would reduce the
problem -- and this is in fact the case. So while there is no simple
solution to avoid this problem entirely, if you add additional washing
steps (particularly, low-spin steps that separate cells from debris),
you will significantly reduce the incidence of this problem. (That's
partial solution number 1).
Practically speaking, though, you can try to ignore it. This issue
never results in increased background -- so the background gates you
establish on FMO's (or whatever control you have) are still valid, no
matter how far below this gate the "negative" cells wander. And while
theoretically this problem can diminish the measured fluorescence of
positive cells (and in fact we've seen that it does so), the
diminution is fairly small and would only affect some dim cells -- and
given that the staining is very bright to see this problem at all, it
probably isn't a real issue. (That's partial solution number 2).
Unfortunately, this does lead to some ugly displays, where the
"negative" cells are far more negative than they should be. There's
nothing you can do about that, except to repeat the mantra that "all
cells below measurable fluorescence are the same" -- i.e., there are
no subpopulations of negative cells!
On Jun 18, 2009, at 4:47 PM, HSCI - CRM Flow Cytometry Core Facility
> Dear flow cytometry experts,
> Several of our researchers are having the same problem. Voltages are
> set based on an unstained sample. Then a CD150 sample is run. The
> negative population appears MORE NEGATIVE than anything in the
> unstained sample.
> We have seen this on FACSAriaI, AriaII, and LSRII. We have seen this
> with CD150-PE as well as CD150-PECy7. We are hopelessly confused. How
> is this even possible? Has anyone out there seen anything like this?
> I'm attaching a picture (tif file) of what I'm talking about.
> Thank you,
> Kat Folz-Donahue
> Flow Cytometry Core
> Harvard Stem Cell Institute | Center for Regenerative Medicine
> Massachusetts General Hospital
> kfolzdonahue at partners.org - 617.643.0790
> Cytometry mailing list
> Cytometry at lists.purdue.edu
More information about the Cytometry