Re: Paul's Tutorial on FL1, FL2, etc.

From: Howard Shapiro (hms@shapirolab.com)
Date: Wed Dec 19 2001 - 07:27:03 EST


Paul Robinson wrote-


"If you want to discuss fluorescence signals, it is appropriate to
use a terminology that is interpretable in a scientific manner. Here are my
suggestions, and these are the ones that we require for those who contribute to
Current Protocols in Cytometry, the manual that I hope all of you have in
your labs!!!

To express a fluorescence signal on any machine we request the fluorochrome
(example follows), and the center of  the wavelength band. e.g. FITC-525
nm, PE-575
nm. If there is an antibody attached, we would suggest it be written
CD4-FITC-525 nm.
Of course, there are many times when we might simply measure a frequency
band, so
FITC may not be appropriate. In such case it might be "green
fluorescence-515 nm" or
similar. If you are using a system with fixed filters that might be "green
fluorescence-
525 nm". On flexible systems where all sorts of filter changes can be made,
then you
obviously must express the specifications of your detection in detail. If
you don't know
what the exact specifications are for each PMT, you shoudl find out, print
them out
and paste them on the front of the machine.

This is not perfect, as it does not necessarily indicate the width of the
band, etc.
However, all of these can be stated in figure legends. If you want to
indicate where on
your cytometer this is measured from, you can add the PMT designation as
well. e.g.
CD4-FITC-523 nm-FL1

No system is perfect, but the system I have suggested is totally
reproducible by any
one with any machine. There are many minor modifications to the above that are
perfectly acceptable.

I take this issue quite seriously, and if you are unlucky enough to have your
manuscripts sent to me for review, I won't accept FL1, FL2, etc. I just
return the M/S
and suggest that the author write in an acceptable scientific format. That
might sound
tough, but in fact there are few if any journals that will accept figure
references that
lack units. That is in fact what we constantly see in flow cytometry. We
need to raise
our standards to what the rest of the community requires."

I would take what Paul wrote a little bit further.

What the *instrument* initially measures are *optical* properties of cells
or other particles:

For "Small angle light scattering", aka forward scatter, FALS, etc., the
measurement depends on the wavelength(s) used, the smallest and largest
included angles, and the shape and orientation of the blocker bar, if used.

For "Orthogonal light scattering", aka RALS, side scatter, 90 degree
scatter, etc., the measurement also depends on the wavelength(s) used; it
is less dependent on angular range than the small angle scattering
measurement, but more dependent on polarization (in fact, one can measure
depolarized orthogonal scatter as a separate parameter; it is useful, for
example, for identifying eosinophils).

Fluorescence measurements are also polarization dependent, and measurements
of fluorescence in planes of polarization perpendicular and parallel to the
excitation beam provide information about the mobility of fluorescent
materials in or on cells.  However, we usually tend to neglect polarization
in specifying how scatter and measurements were made, sometimes (see Asbury
et al in the June 2000 Cytometry) at our peril.
Obviously, one needs to specify the excitation wavelength (or range in
instruments not using strongly monochromatic sources, e.g., lasers), and
the emission passband.  This can be done most precisely when band pass
filters, characterizable by center wavelength and bandwidth, are used; if
you use a long pass filter, the detector bandwidth may be more important
than filter characteristics in limiting response at the long end, as most
people who work with Cy7 tandems have noted by now.

But that describes the raw measurements we make, *not* how we should label
the axes of our data displays.  On the scatter side, I don't have a problem
with FALS, RALS, or forward and side scatter, but the FALS signal should
*not* be labeled "cell size" unless the experimenter can produce a
calibration curve (for example, the B-D FACScan, according to Current
Protocols, produces a bigger signal from a 5 um bead than from a 5.5 um bead).

And the fluorescence channels aren't what the raw measurements are, in most
cases, because we *compensate* them.  The machine, using 488 nm excitation,
may, for example, measure 525/15 nm fluorescence, 575/20 nm fluorescence,
and 670/20 nm fluorescence (n1/n2 here give center frequency and
bandwidth).  We use compensation to get the intensity of fluorescence from,
for example, fluorescein (not "FITC", or fluorescein isothiocyanate; that
is what the antibody was conjugated with and there ought not to be any in
the sample), phycoerythrin (PE), and PE-Cy5 tandem conjugate labels.  It is
acceptable to specify either the fluorescent label, or, more scientifically
interesting in terms of the biological characteristics of the cells, the
antibody or other reagent to which it was bound in the label(legend) of a
display.  It is inappropriate to label an axis with the optical
characteristics of the channel if the data are compensated.

Of course, everybody who reads this list does the flow and the compensation
correctly, so we can be believed if we neglect to tell people which machine
we used, what the wavelengths were, and which antibody had which
fluorescent label, and just put the antibody names on the axes.  Well,
maybe not, but, in an ideal world, that's what we'd want.  And for
non-antibody measurements, we ought to label axes as "DNA content", "RNA
content", etc.  Membrane potential, pH, calcium, etc. are a bit dicy if
they are not calibrated, but, in an ideal world, they are, *or* there is
some disclaimer about the imprecise nature of the measurement somewhere in
the paper.

More to come in the 4th Edition.

-Howard



This archive was generated by hypermail 2b29 : Sun Jan 05 2003 - 19:01:44 EST