Yes, I think you have the background subtraction *formula* right.
Begin with the assumption that the isotype control spillover into
the positive region is due to a random selection of the cells which
will be distributed after staining in the same proportions as the
rest. That is, instead of subtracting *all* the spillover, just take
off an appropriate fraction:
y = x - c(1 - y/100)
y = actual % +ve
x = measured % +ve
c = measured % +ve for isotype control
which rearranges to your equation. However, can I suggest two extra
considerations:
1: Is the assumption correct? Are the spillover cells random or are
they predominately one cell type (large, sticky, autofluorescent..)?
Thus, maybe all or none will be specifically stained in the positive
sample (depending on the actual stain and the actual sample).
2: You give the example of a positive sample 100% brightly stained.
If "brightly" means way above the cutoff point that gave the value
"c" for the isotype control, you may get another answer using a higher
cutoff point for both control and sample.
(I don't have the answers, just the questions)
Regards, | | < Frank: battye@wehi.edu.au
\__/ <<<<< The Walter & Eliza Hall Institute
---------------!!<<<<<<<< Post Office, Royal Melbourne Hospital
/!!\ <<<<< Victoria 3050, Australia
o !! \ < ph: 61_3_9345 2541, fax: 61_3_9347 0852