Re: Problems seeing Cy-5 and PE-Cy5

From: Howard Shapiro (hms@shapirolab.com)
Date: Mon Mar 20 2000 - 15:34:51 EST


Howard Robinson writes:



>Problem:The signals in the Cy5 and PE-Cy5 channels [on a Cytomation MoFlo,
>equipped with a Melles Griot 35mW Helium Neon laser (Laser #2)] are very
>low.  On unstained cells
>the signal only occupies 20% to 30% of the first decade on the display.
>Essentially all the events in these channels are cramped down near the zero
>channel.  This is a real problem with dual parameter histograms such as FITC
>vs Cy-5.  When the Cy-5 negative population is so low, the FITC (pos)/ Cy-5
>(neg) population is difficult to discern.
>
>In discussing this problem with Cytomation, increasing PMT voltages to push
>out the negative population causes thermeonic emissions in theses channels.
>So, we are faced with either negative populations that are too low to view
>data or we are faced with creating an erroneous population.
>
>Question:
>How can we move out the negative population without thermeonic emissions
>creating a ghost population?

Unstained cells shouldn't be very fluorescent in the emission range
associated with Cy5 (and PE-Cy5), so one would not expect them to occupy
more than a fraction of the first decade on the display.  Essentially,
there isn't a signal to amplify, so the only way you can move the negative
population where you like it is to increase the gain, and what you see will
be noise due to thermionic emission and other electronic sources.  But what
do you mean by a "ghost population"?  If you mean that some of the negative
cells come off the base line and some don't, that effect is typically an
artifact of fluorescence compensation.  The only way to make it go away on
the display is to add random noise to the signal or random numbers to the
digitized data.  Just as is the case when you turn the PMT gain way up, you
are making the data look better by adding noise.

On a Cy5 channel, one can increase the apparent signal from unstained cells
by using an emission filter which lets some scatter through - one would
generally prefer to avoid doing this, because, again, you're adding noise,
and this will decrease the separation between the Cy5-positive and
Cy5-negative cells (as will all the other methods of adding noise).

As long as your gate on a 2-D histogram can include the axes, there
shouldn't be a problem with analyzing the data as you get them; if that's a
problem, gate out the fluorescein-positive population on a 2-D histogram of
green fluorescence vs. scatter and do a 1-D histogram of Cy5 to find
positives and negatives.

-Howard Shapiro



This archive was generated by hypermail 2b29 : Wed Apr 03 2002 - 11:55:38 EST