Re: background subtraction

Ben_Verwer@bdis.com
Tue, 07 Nov 95 18:32:51

I never thought of this but I like it. One assumption you have to mention though
is that any cell which stains specifically has to be lifted above the control
marker. One can imagine that cells pick up stain but not enough to go above the
cutoff, but enough to make a shift in the histogram below the cutoff. In that
case the formula obviously does not hold and one would have to use some more
refined mechanism like estimating the parameters of the underlying
distributions.

I don't think your formula is obvious, although it is suprisingly simple and
although I have seen the formula before (in glare correction in microscopy,
where pixels - due to stray light - pick up light when they are dark, but loose
light to neighbouring pixels by the same mechanism as well).

_______________________________________________________________________________
Subject: background subtraction
From: Alice.L.Givan@dartmouth.edu (Alice L. Givan) at INTERNET
Date: 11/7/95 10:13

Here's an old problem: What do people do about subtraction of isotype control
values from flow determination of the so-called "% positive" population? In
most lab with which I am acquainted, if the isotype control marker is set at
3%, then 3% is subtracted from every subsequent sample. Although I have never
heard it discussed in public, I have known for a long time that this approach
is obviously incorrect -- as can be seen if one imagines a sample with clearly
ALL the cells brightly stained. By subtracting the control, it would then
appear that only 97% of the cells were stained. The problem with this
"classical" approach is that we should be subtracting 3% of the true negative
cells only because those are the ones that are mistakenly contributing to the
determination of the positive population.

When I follow this through algebraically, I come up with the following
formula:

y=(x-c)/(1-c/100)

where y is the actual % positive
x is the measured % positive
c is the % positive set for the isotype control

Am I wrong about this (or is it so obvious that everyone else does it already)?
Does someone have the patience to check my derivation?
Thanks.

Alice L. Givan
Englert Cell Analysis Laboratory
Dartmouth Medical School
Lebanon, NH 03756
tel 603-650-7661 or 7907
fax 603-650-6130


Home Page Table of Contents Sponsors Web Sites
CD ROM Vol 2 was produced by staff at the Purdue University Cytometry Laboratories and distributed free of charge as an educational service to the cytometry community. If you have any comments please direct them to Dr. J. Paul Robinson, Professor & Director, PUCL, Purdue University, West Lafayette, IN 47907. Phone:(317) 494-0757; FAX (317) 494-0517; Web http://www.cyto.purdue.edu EMAIL robinson@flowcyt.cyto.purdue.edu