Re: 4 color staining and Argon laser

From: woodbl (woodbl@u.washington.edu)
Date: Fri Jan 22 1999 - 13:54:42 EST


Hi,

Yes, we have much experience with the system you describe.  It took us a
while to figure out how to deal with it, but it works well for us now.  The
issues for us were:

1.  Each ECD antibody we have used (Coulter) has a different amount of PE
leakage, so the amount of compensation required is different for each ECD
antibody.  CD8 has had the most leakage (which we initially tried), and
results in significant overcompensation of the PE signal when other ECD
antibodies are used.  CD45 has tended to be next, followed by CD3.  CD19 has
had the least leakage.  CD45, CD3 and CD19 are close enough that one can
approximate the comps using any one of them for compensation.  Ideally one
needs to have a separate compensation matrix set up for each ECD antibody
used.

2.  Use median values, not arithmetic or geometric means, for compensation. 
The fluorescence distributions of the population, in particular the negative
population, is not normally distributed and incorrect compensation will
result.  Unfortunately, the Coulter automatic compensation software uses a
mean for its calculations and results in suboptimal compensation, most
noticeable as marked PE overcompensation.  If one attempts to use medians,
often the median for negative populations in the higher wavelength PMTs lies
very close to the axis and makes accurate determination of the comps
difficult.  The best solution for us has been to use off-line software
compensation with WinList.  The resulting comps are consistent and reliable,
and the amount of time required to perform the compensation is dramatically
reduced.  However, one does have to run the samples through the instrument
uncompensated.  Since we do all our data analysis off-line anyway, this is
not a problem.

3.  Do not expect to use quadstats to analyze the data.  Due to the
significant spectral overlap of ECD with both PE and PE-Cy5, the true
negative does not lie parallel with the axis (unlike orthogonal quadstats). 
This effect can be quite marked and can best be seen on histograms
displaying PE vs. ECD and ECD vs. PC5.

Hope this helps,

Brent Wood MD PhD
Associate Director of Hematology Laboratory     
University of Washington Medical Center         
EM: woodbl@u.washington.edu  Phone: (206) 548-6199  Fax: (206) 548-6189

----------
>From: "robert gniadecki" <rgniadecki@hotmail.com>
>To: Cytometry Mailing List <cytometry@flowcyt.cyto.purdue.edu>
>Subject: 4 color staining and Argon laser
>Date: Fri, Jan 22, 1999, 6:57 AM
>

>
>Hello!
>We are planning immunophenotyping with 4 colours: FITC, PE, Pe-Texas red 
>(ECD) and Pe-Cy5. We have however experienced some severe problems with 
>compensation between PE and PE-TR.
>
>Could you share your experience with me? Is that fluorochrome 
>combination possible to use?
>
>We have 4 photomultipliers and a 488 nm argon laser.
>Thanks, 
>
>Robert Gniadecki
>Bispebjerg hospital
>Copenhagen, Denmark
>
>______________________________________________________
>Get Your Private, Free Email at http://www.hotmail.com
>



This archive was generated by hypermail 2b29 : Wed Apr 03 2002 - 11:52:58 EST