 |
 |
 |
 |
Several factors influence bead cv measurements. One is sample flow rate, which
dstermines sample core size in the jet (or cuvette). The smaller the sample
core, the more likely each bead will traverse the same path in the jet through
the excitation, and the more uniform the measurements should be. So, for most
flow instruments, one way to achieve better cv's is to slow down the sample
flow.
The illumination and collection optical geometry also can have a great effect on
the measured cv; most people will have to assume that it is a given, since they
are not going to redisgn their setup. However, insuring the cleanliness of all
these parts can result in a stronger signal which would reduce measurement
error.
Lasers (if used for excitation) are another place where error can be reduced. If
you have control over the laser aperture, adjusting for a gausian profile is
important. Also, in general, the higher the excitation power, the less
measurement error there will be due to statistical variation in the signal.
However, with real dyes there is a trade-off between higher autofluorescence of
cells (depending on excitation/emission wavelengths and cell type) and dye
saturation (depending on dye type).
Lastly, here is a method we have used to estimate both the number of photons
measured for a given bead (very useful for comparing machines) and an estimate
of the intrinsic cv of the signal source (which includes bead cv).
1) Under normal operating conditions, measure bead signal strength and
cv.
2) Using various neutral density filters in the light collection path,
measure the cv and signal levels at the same laser, flow rate, and PMT settings
as (1). You might have to increase the gain of the linear amp used in the
measurements in order to do this.
3) Plot the data as 1/(signal level) on X and CV**2 on Y. The
y-intercept is an estimate of the cv**2 of the source, which includes the bead
cv, and the slope yields an estimate of the number of photoelectrons the source
is producing, using the formula:
EVENTS = 1/(cv**2 - CVo**2),
where CVo**2 is the y-intercept on the graph.
Hope this helps.
-Marty Bigos
Stanford Shared FACS Facility
|
|
|
|