re: negative gating in rare event analysis

From: John_Dunne@BDIS.Com
Date: Wed Feb 13 2002 - 17:20:32 EST


Hello Michal,

Negative gating can be very useful in multicolor applications, in that it
capitalizes on the propensity of nonspecific staining to occur with all the
antibodies in your staining cocktail (this assumes that your antibodies are all
near their respective optimal concentration, and that none of them are
aggregated).  The general notion that seems descriptive is that most cell preps
have some cell debris, and that this debris can stick to cells and to
antibodies.  To the extent that this problem contributes to your "background,"
you can magically gate it away by focusing your attention on the cells that
didn't bind one of the antibodies in your cocktail, a "negative gate."

Sometimes referred to as a "dump channel," lots of investigators simply add an
irrelevant antibody with a fluorchrome that is not being used in the important
part of the assay.  Like the specific antibodies, this irrelevant antibody will
bind to the debris, and applying a negative gate on this irrelevant antibody
will filter out cells which have this kind of background staining, and sometimes
pretty miraculously turn a mess into a publication.

The most common implementation of a related strategy is to gate out dead cells
which stain with propidium iodide.  Dead and damaged cells are notorious for
nonspecific staining, and a PI-negative gate is very generally used.  Even in
applications where the PI stain will contaminate other important red signals,
the PI+ cells are much brighter than most immunofluorescent signals, and you can
use a PI-bright gate to generate meaningful data about the dimmer, healthy
antibody-PE positive cells.

The utility of these tricks with tetramer staining makes sense and ought to
work, but it might be worth thinking about.  You're using labeling reagents with
pretty different protein structure and molecular weight, maybe different kinds
of irrelevant binding features, storage buffers, etc.  Note that any gating
strategy can be a complicated filter, and unless you're pretty sure you
understand and control for subtle effects, it's certainly caveat emptor.  I
don't think it's really very different from filtering a clumpy sample with 70um
mesh...you never really know what kind of bias you have introduced to the
sample, so that interpretations of the frequency of a given population can be
complicated.  The best advice is certainly to use sample prep strategies that
minimize debris in the first place.  JD>



This archive was generated by hypermail 2b29 : Wed Apr 03 2002 - 11:59:24 EST