Here are the best reponses that I received to this recently inquired
subject. I have include the email addresses of the responders in case
anyone wishes to contact them directly for additional information.
As you will see, I received many varied and most helpful responses-see why
I love this user group
Janet
From: Ian_DIMMICK@europe.bd.com
Subject: Re: meaning of FL2-area and FL-2 width
Dear Janet,
(Embedded image moved to file: pic00153.pcx)(Embedded image moved
to file: pic00292.pcx)
>From the above two diagrams you can see the generation of a pulse in
diagram 1 as a cell passes the interrogation point where the laser is
focussed on your cell nuclei, in the second diagram the pulse can be
processed in various ways
The height of the pulse is proportional to the intensity of the
fluorescence signal (in your case PI) and in part due to the shape of the
laser interrogation spot.
The width of the pulse is proportional to the time taken for the cell
nuclei to traverse the interrogation point of the laser (so called time of
flight)
The area of the pulse is calculates from the integral fluorescence (area)
that is cross hatched in diagram 2
The reason for using these two parameters(area v width) is primarily
to discriminate single cells passing through the flow cell from cells that
may have no spacial separation when passing through the flow cell ie
doublets.Doublets will give an increased width measurement (26% increase
for every doubling of volume of a cell based on a spherical arithmetic
model), The area measurement will also increase however when analysed the
histogram of width v Area will enable you to gate out any
doublets(disproportionate width to Area) and therefore avoid mis
interpereting a Go doublet for a G2m event
Forgot to mention you can use height , however it is felt that you get
better discrimination using width and area , but it realy depends on cell
size, shape and laser beam geometry .
!!!Unfortunately I was unable to open the diagrams that Ian sent me so
contact him directly for copies of them!!!
From: Alice.L.Givan@Dartmouth.EDU (Alice L. Givan)
Subject: Re: meaning of FL2-area and FL-2 width
Hello Janet,
When a cell goes through the laser beam, it emits fluorescent light. That
emission of light takes place over time. The emission starts as soon as
the leading edge of the cell enters the laser beam and the emission
continues until the trailing edge of the cell leaves the laser beam. So
the light signal has a certain shape over time (this is actually like an
oscilloscope trace). If you plot out that light signal over time you will
get a plot that has a certain height (the height is the maximum amount of
fluorescence that is given out an any one instant of time); the light
signal will also have an area (this is the total integrated amount of
fluorescence that is given out from the entrance of the cell into the laser
beam until the final moment when the cell has entirely left the laser
beam); and it will also have a width (this is the total amount of time
that it takes the cell to traverse the laser beam).
Since we are usually talking about cells that are larger in diameter than
the width of the laser beam, then only the area of the signal (that is,
the total integrated fluorescence) will be proportional to the total DNA
content. If you use the height of the signal, it will not be proportional
to the total DNA content if the cell is elongated and doesn't ever fall
entirely in the laser beam at any one time. So using the area of the
signal is important if you want to compare cells that might be in G1 with
those in G2 of the cell cycle (these cells could be of different sizes and
would, therefore, not be equally contained in the laser beam at any one
time).
In addition, the width of the signal can be used to distinguish G2 cells
from clumps. This is because a clump of two cells takes longer to go
through the laser beam than does a single G2 cell. So the width of the
signal (actually the length of time over which the signal is emitted) is
important in cell cycle analysis to rule out aggregates of two G1 cells
that can masquerade as a single G2 cell because two G1 cells have the same
DNA content as one G2 cell.
If you would like, I could mail you a diagram from my book. Send me a fax
number, if this would help.
Alice
From: Andy Johnson <andy@brc.ubc.ca>
Subject: Re: meaning of FL2-area and FL-2 width
Wow Janet.....you should get alot of replies from this question.
It is a very easy question to answer and an area that is normally missed
unless you are trained by somebody.
>I know this is going to sound like a stupid questions but we(me and several
>post docs) are currently involved in a heated debate as to the nature and
>meaning of FL2-area and FL2-width in connection with the measurement of
>DNA content(PI stained) of ethanol fixed cat lymphocytes.
It is critical to use as least one pulse processing parameter...either Area
or width, or both if you so wish.
When a cell passes through a laser it causes a pulse. If we concentrate on
FL-2 This pulse has a height (due to fluorescence) a width (due to length
of time it takes to pass through the laser = size) and an area (height x
width/2).
These extra parameters allow you to easily distinguish a doublet (two
cells) in G1 from a single G2 cell. Both have the same FL-2H
(fluorescence), but the doublet will have a larger width or area as the two
cells are physically larger. Without this you wouldn't be able to tell as
they both have roughly the same amount of DNA stained.
If you search "flow cytometry" and "cell cycle analysis" in www.google.com
then you should find some links to pages with pictures and more details.
Andy
Janet -- The reason people use fluorescent area and fluorescence pulse width
measurements is to try to remove doublets, triplets, etc. from their data
prior to mathematical analysis for cell cycle information. Some instruments
have the ability to make some or all of these measurements and some don't.
Beyond that, some instruments make relatively crude measurements while
others have considerably higher resolution. So whether you use a "hardware"
approach using these electronic processing methods, or a software that
corrects for doublets based on triplets, or other algorithms, depends on
what instrument you have, what application you are running (including what
the relative size of your cells compared to the laser beam width which is an
important factor, or if you have high-resolution, real-time laser beam
subtraction capability which greatly changes the problem).
Your question really has a number of factors that make it difficult to
give a very brief answer. I could, but it might be wrong for your (or
another ListServ reader) circumstances. So rather than either giving a
simplistic answer or, on the other extreme, a very lengthy answer with some
of the nuances, I will refer you to a paper that an NIH group and I
published last year. We not only discuss your question but explain why you
get different quality of results on different instruments using these
hardware approaches on the signals as they come in or alternative software
approaches on the data after acquisition. So I hope you find this article
helpful and informative. If not come back for more...I've been doing these
measurements on a wide variety of cell types for more than 20 years and have
learned much in the school of hard knocks!
Wersto, R.P., Chrest, F.J., Leary, J.F., Morris, C., Stetler-Stevenson, M.,
Gabrielson, E. "Doublet Discrimination in DNA Cell Cycle Analysis"
Cytometry 46(5): 296-306, 2001.
- -Jim Leary
James F. Leary, Ph.D.
Chief, Molecular Cytometry Unit, Div. Infectious Diseases
Professor of Internal Medicine, Pathology, Microbiology &
Immunology, Biophysics, Human Biological Chemistry &
Genetics, and Biomedical Engineering
4.216 Mary Moody Northen Pavilion - Route 0435
University of Texas Medical Branch
301 University Blvd.
Galveston, Texas 77555-0435
Tel: 409-747-0547; Fax: 409-747-0550
From: Larry Arnold <lwarma@med.unc.edu>
Subject: Re: meaning of FL2-area and FL-2 width
Mime-Version: 1.0
Janet
Parameter height, area, and width can be visualized by looking at the curve
(pulse) one sees if looking at (or envision) an oscilloscope tracing of a
particle transiting the laser beam. The height is the height of the pulse,
the area is the area under the curve and the width is the width of the
pulse at some threshold level (for linear signals only). The area always
provides the total fluorescence of the cell so for DNA measurement this is
what we want. If the laser beam is wider than the cell the height and area
are proportional. If the laser beam width is less than the cell diameter
(slit scanning) then this is not true and you need to use area to get the
total fluorescence. Envision 2 identical cells with the same number of
fluorescent molecules attached. In one cell the fluorochrome is
concentrated in one small area whereas with the other cell the fluorochrome
is evenly distributed over the entire cell. In the first cell (using a
narrow laser beam and disregarding any fluorochrome interactions which may
quench some fluorescence) the height of the fluorescence signal will be
very large but the width will be very narrow. In the second cell the
height will be much less but the width will be greater. Using height one
would conclude that the first cell has more fluorescence which of course is
untrue as we set up the experiment. Using the area both cells will be
evaluated as having the same fluorescence. Now back to your situation
where the laser beam is almost certainly wider than the cell (nucleus)
diameter. Here width for a single cell is actually the width of the laser
beam. Envision a single G0 cell transiting the laser compared to a G0
doublet (with the doublet longest axis perpendicular to the laser
beam). The height of the doublet (in a FACScan) will be slightly greater
than that of a singlet (if the laser beam diameter was the size of a
nucleus then the height would be exactly the same since no more than one
nucleus worth of PI could be in the laser beam at any one time). Note that
these conditions change as the size of the cell (nucleus in the case of
DNA) changes since the laser beam diameter is fixed. However the width of
the doublet compared to the singlet will be much greater (i.e. up to
twice). The area of the G0 doublet will also be twice that of the
singlet. Without any doublet detection/correction you will measure this G0
doublet as a single G2/M cell - a significant error. By plotting the width
versus the area you can visualize the G0 doublets (and higher order
complexes) from the singlet G2/M cells and gate them out. (You can also
use height vs area but width vs area on the FACScan/Calibur seems more
sensitive). This is hardware doublet discrimination. Software methods
(e.g. ModFit) calculate (and then subtract) the number of G0 doublets by
counting the triplets, quadruplets etc. However, this often can not be
used depending on the cell populations present (for example with
endoreduplicating cells). Note that the doublet G0 population width is a
smear (on a histogram) rather than a tight population. This is because not
all doublets will transit the laser beam with their longest axis
perpendicular to the laser. As the angle between the longest axis and the
laser beam becomes less than 90 degrees then the width decreases. If the
doublet long axis is parallel (or close to parallel) to the laser beam the
doublet is not detected. Thus, the fluidics system needs to be operating
properly - i.e on the FACScan running on the low fluidic setting (this is
critical for the best CV also as well as separating independent doublets (2
cells not connected but in the laser at the same time)).
Hope this helps.
Larry
From: "Nancy Harvey" <nharvey@microbio.umsmed.edu>
Subject: Re: meaning of FL2-area and FL-2 width
BD's manual for Pulse Processing (Vantage, FACStar cytometers) gives a
pretty good explanation.
You might want to talk with BD's technical support people.
Nan Harvey
RE: meaning of FL2-area and FL-2 width
Dear Janet,
When looking at the pulse of fluorescent light emitted from PI (7-AAD,
etc.) after laser excitation the two primary parameters for discriminating
the cells is peak fluorescent height and the area under the peak (integral)
or time of flight as it passes through the laser beam.
http://nucleus.immunol.washington.edu/Research_facilities/Apps/logscale.
html
Here is a web page from our site that demonstrates how this is used to
separate aggregates from single cells.
http://sciencepark.mdanderson.org/flow/files/Aggregates.html
Hope this helps.
Sincere regards,
Kent Claypool
Facility Core Manager
Cell & Tissue Analysis
Department of Carcinogenesis
Center for Research on Environmental Disease
University of Texas
MD Anderson Cancer Center
Science Park Research Division
Smithville, TX 78957
512.237.9427 Lab
512.237.2444 Fax
http://sciencepark.mdanderson.org/flow
From: "Pizzo,Eugene" <Pizzo@NSO1.UCHC.EDU>
Subject: RE: meaning of FL2-area and FL-2 width
Janet,
The best and clearest explanation I've seen for this, and it really
is quite straightforward, can be found in Ormerod's 4th chapter on
DNA analysis in "Flow Cytometry, A practical Approach" IRL press
ISBN:0-19-963052-6
Gene Pizzo/UCONN Health
Suffice to say width measurement gives you a time component
through the laser for the fluorescence measurement which enables you to
distinguish cells which have a true G-2 level of fluorescence from
cells which are aggregates and thus may only appear to have higher
levels of fluorescence because they take longer to pass through the laser.
Area as I see it can simply be viewed as a more resolved height
measurement in which the total fluorescence(integrated peak) rather than the
peak
of the fluorescence is taken as a final value. A value which is
only necessary or useful when high levels of discrimination of
absolute fluorescence levels is neeeded as in DNA analysis.
From: "Melendez, Johana" <MelendJ@moffitt.usf.edu>
Subject: RE: meaning of FL2-area and FL-2 width
For acurate analysis, the instrument must be able to discriminate between
two G0/G1 cells stuck together (a doublet) from a true G2 + M cell. Both
would appear tetraploid (4N). Doublet discrimination is a function of pulse
processing. A pulse is the signal generated as a fluorescent cell moves
through the laser. There are 3 key features to pulses. The first is the
pulse height: the maximun fluorescence intensity detected as the center of
the cell moves through the laser. Second is the pulse width: the the
relative time it takes for the cell to pass through the laser. Third, the
pulse area is an integration of the pulse height and width which describes
the fluorescent properties of a cell as a whole. Doublets and triplets can
be identified by alterations in their pulse profile (the shape of the pulse
poduced).
In summary, we must use this parameters to distinguish doublets, or
aggregates from single cells. So we don't confuse them.
(To better visualize this go to the international Society for Analytical
Cytology website).
From: "Dennis J. Young" <djyoung@ucsd.edu>
Subject: Re: meaning of FL2-area and FL-2 width
Mime-Version: 1.0
I think it is strange that there would be a heated debate.
The quantity of DNA is best estimated by the amount of fluorescence. The
height, area and width parameters refer to the Gaussian-like shape of the
multiple-photon count signal that a cell makes as it passes through the
light source. The longer it takes to pass, the larger the particle, so the
width of the signal helps identify aggregates. On the FACSCalibur, the peak
vs area is not as good at discriminating peaks, but can be used. The total
amount of fluorescence (DNA) is better measured by the area of the signal,
rather the peak (height).
Howard Shapiro has provided FAR more detail on this in the archives and in
his Practical Cytometry.
Perhaps you did not find anything useful in the archives because the titles
of the posts are not descriptive. The answers are buried in the text. There
was a post by Joe Trotter when he was still at Scripp's explaining the
difference in using FL-2 or FL-3.
I use FL2-A vs FL2-W to gate clumps and at the same time use FL3-H in log
mode to measure sub-G1 when appropriate.
From: Eric Van Buren <aa9080@wayne.edu>
Subject: Re: meaning of FL2-area and FL-2 width
Janet Dow,
Although this is easier to explain with a figure, I will try to briefly
explain these signals with text. (Let me know if your email can accept a
100-200k jpeg, and I will send you a copy of a good figure from BD.)
As the cells pass through the laser beam, they produce a signal (from the
FL2 photomultiplier tube, for example) versus time. This signal is usually
a "bell curve" (gaussian) shaped peak, with minimal (baseline) signal before
and after the cell is illuminated, and a maximum signal when the cell is in
the center of the laser beam.
The height of this signal peak -- the maximum signal when the cell is in the
center of the laser beam -- is the easiest measurement to make electronically.
This is your FL2-H (FL2-Height) parameter. It represents the maximum
fluorescence of each cell. For immunofluorescence experiments, FLx-Height
measurements are adequate, since the difference between "positves" and
"negatives" is typically 100- to 1000-fold.
Possibly a more accurate measurement would be to measure the total fluorescence
of each cell. This can be pictured as the "area under the curve" or integral
of the signal versus time peak. This is your FL2-A (FL2-Area) parameter. For
DNA content experiments, the (PI or other) fluorescence is proportional to DNA
content, so it makes sense to measure the total fluorescence, as this is
proportional to the total DNA content. Maximum fluorescence (FL2-Height) also
increases with DNA content, but total fluorescence (FL2-Area) is a better
measure of DNA content. Theoretically, the ratio of FL2-Area means for G2 to
G1 is exactly 2, whereas the ratio of FL2-Height means will always be less
than 2. [On Coulter cytometers, doublet discrimination is done with these
two parameters, but Coulter calls the Area parameter "Integral" and the Height
parameter "Peak" (short for peak height), so it is Peak versus Integral.]
Lastly, the electronics can measure the width of the signal peak. Since this
is simply the distance along the x-axis, it is equal to the time it takes
each cell to pass through the laser beam, and since larger cells take more
time than smaller cells, it is also a measurement of cell size. This is your
FL2-W (FL2-Width) parameter. Doublet discrimination is based on the idea
that one G2 cell has the same amount of DNA as a G1-G1 doublet (FL2-Area),
but the G1-G1 doublet is larger than one G2 cell (FL2-Width). This is only
slightly complicated by the fact that as cells progress from G1 through S to
G2 they get larger. That is why when viewing FL2-Area versus FL2-Width, the
line from G1 to G2 is slanted, and not a straight vertical. [On BD cytometers,
there is better discrimination of doublets when viewing Area versus Width --
as opposed to Height versus Area on Coulter cytometers -- and that is why
these two parameters were chosen.]
There is one caveat to using DDM on BD cytometers that BD does not always
stress enough. The DDM electronics -- especially those concerned with measuring
the Width parameter -- only work properly for signals above channel 50 (on a
1024-channel scale) when measuring the Height parameter with linear ampli-
fication and gain set to one. So, even though FL2-Height is not important
during data analysis, it is quite important when setting up acquisition. If
the dimmest population of interest on FL2-Area is in channel 100 or 200 or
higher, typically its FL2-Height will be above channel 50, and this is
probably why BD does not fuss over this point.
Another issue that has been discussed on the Purdue email list is the
occurance of cells with FL2-Width in channel zero. Evidently the DDM elec-
tronics will fail to produce the proper FL2-Width measurement -- apparently
randomly -- for certain sample preps and not others. I choose to ignore those
events with a FL2-Width of zero, which is what normally happens if you draw
your singlet region the way BD shows you (away from the FL2-Width axis).
The author of ModFit insists that an ungated FL2-Area histogram modeled to
include aggregates is more accurate than a DDM-gated FL2-Area histogram
modeled without aggregates. Anectodally, the few times I have looked at
this issue I was more impressed with the DDM results.
In summary, FL2-Area is a measurement of total PI fluorescence, which is
proportional to total DNA content, and has better "linearity" than FL2-
Height, which is a measurement of maximum PI fluorescence. FL2-Width is
a measurement of the "time of flight" of each cell, which is related to
cell size. A plot of FL2-Area versus FL2-Width allows for the discrimination
of aggregates from singlets, and the resultant gated FL2-Area histogram
will be free of aggregates for cell cycle analysis.
Not as brief as I wanted it to be, but I hope you find my answer helpful.
Good luck,
Eric
From: "Wersto, Robert (NIA/IRP)" <WerstoR@grc.nia.nih.gov>
Subject: RE: Area vs width
Date: Wed, 9 Jan 2002 13:44:37 -0500
Hi:
Regarding your question on the Purdue website, you might be interested in:
Wersto RP, Chrest FJ, Leary JF, Morris C, Stetler-Stevenson MA, Gabrielson
E.
Doublet discrimination in DNA cell-cycle analysis.
Cytometry. 2001 Oct 15;46(5):296-306.
Janet Dow
Research Technician and Manager
Flow Cytometry Facility
North Carolina State College of Veterinary Medicine
Room C-314
Raleigh, NC 27606
(919)513-6364
This archive was generated by hypermail 2b29 : Sun Jan 05 2003 - 19:25:55 EST