These questions come up regularly and emphasise the importance of basic
understanding of cytometry principles such as compensation. Your
experiment suffers from "very strong green fluorescence signal" syndrome
as explained at the bottom of the message.
There are a couple of websites to look up. Applied Cytometry Systems had
a nice short website on the subject but probably lost it in the
com-ercialisation of their webpages. Mario's website is probably the
most comprehensive at http://drmr.com/compensation/index.html.
Below my virtual-picture explanation of the "very strong green
fluorescence signal" syndrome problem slightly edited from the purdue
email archive in 2000:
One of the problems of compensation is that it is indeed not a fixed
value but depends on a number of factors, the dye characteristics and
the response characteristics of the detectors.
The latter is a combination of the spectral response mainly determined
by the filter characteristics (which are quite different between
different instruments even of similar models) and the amplification
settings. For the simplified case of the 'bleed' of FITC into the PE
channel (ignoring other photon sources such as autofluorescence) the
value is constant only for a fixed set-up of both detectors at a given
alignement. Below this is described in a hypothetical example:
The signal received by detector FL1 is 1000 photons which the detector
converts
into 1000mV. Due to the FITC emission and the filter combination, the
FL2
detector will receive 100 photons thus 100mV that can be compensated by
setting
fl2=fl2-10%fl1. A cell that sends 10,000 photos into FL1 (10,000mV) will
therefore send 1000 photons into FL2 (1000mV) which is again compensated
correctly with 10% as above. However, if one changes PMT voltages or
amplification factors things change. If for example the FL2
amplification is
doubled, the overlap of 100 photons will now generate 200mV, thus the
10%
compensation is insufficient or it is to high if FL2 amplification is
reduced.
The inverse is true if only FL1 amplification is doubled. Whilst the
1000
photons will now generate 2000mV and the 100 photons in FL2 will still
only
generate 100mV thus a 10% compensation would lead to a -100mV signal.
In case of a "very bright green fluorescence syndrome" such as CFDA-SE
at too high concentrations one might set FL1 amplification low to get
the signal in scale. If full signal scale would be 10,000 mV, equivalent
to 100,000 photons to be detected, the amplification factor would have
to be 1/10 (or in fact 1/20 to get it in the scale with 5,000mV). At a
10% photon bleed into detector 2 this would still send 10,000 photons
into the FL2 detector, representing 100% compensation (or 200% if the
amplification of FL1 was reduced by a factor of 20).
With the variation between instruments it is not possible to transfer
settings from one instrument to the other. A point to remember here is
also
that filter characteristics are depending on the angle at which the
photons
pass the filter / mirror, a trick usually used to fine-tune optical
filters.
Hope this helps to clarify bits of the discussion
Gerhard
-----Original Message-----
From: joan Kalnitsky [mailto:jkalnits@vt.edu]
Sent: 15 October 2002 19:13
To: cyto-inbox
Subject: CFSE
I have a client who is using CFSE for the first time. She is
also
staining with cd4/cd8, using PE-Texas red and PE-Cy5 TC. When we
started
setting this up we ran a cd4/8 and things looked fine. Then we ran a
CFSE
only tube and it is fluorescing everywhere. We can not comp out the
CFSE
signal from any of the other PMT's. Our PMT setting for CFSE is pretty
low and the peak is set a bit lower than the literature suggests in the
attempt to reduce this "issue".
Does anyone have any input as to why the CFSE is so bright in
all
the other channels?? and if so any suggestions as to how to deal with
this
challenge.
Thanks in advance for all help.
Joan K.
Joan Kalnitsky
Flow Cytometry Lab Supervisor
VMRCVM
(540) 231-4115
FAX 540-231-7367
jkalnits@vt.edu
"It is better to serve than to receive."
B. Borg
This archive was generated by hypermail 2b29 : Sun Jan 05 2003 - 19:26:27 EST