RE : autofluorescence and compensation

Hi Eric,

The simplistic take on it is that compensation is solely dependent on the percentage of
spillover of each fluorochrome between channels, and autofluorescence has no bearing on
this. Having said that, and it sounds like you know, for initial compensation setup you
need to have stained and unstained particles that have identical autoflourescence (see
Mario Roederer's compensation site for a good review). Once this is done however, the
compensation matrix will be valid for particles of all autofluorescence levels...just be
aware that the unstained cells of one type are likely to be in a different position on a
plot to unstained cells of another type. Even if you normally do your compensation using
cells, BDs comp beads are worth their weight in gold for setting up compensation in
populations with heterogeneous AF. If you are working in the mouse, set your viability
stain up on normal mouse spleen and try to only get lymphocytes (live and dead) in the
scatter gate as they have relatively homogeneous AF.

If you have enough free channels in your experiment, think about dedicating one or more
solely to autofluorescence. Instead of it being a hindrance it can actually be useful in
distinguishing different cell populations, particularly in teasing apart different
myeloid lineages. The FITC channel on the calibur is good for this because it can be used
essentially without any compensation most of the time, but you can use other channels if
you compensate the true fluorochromes out using your preferred flavour of
post-acquisition analysis software (Diva is a pain for this but with FlowJo it can be
done relatively easily). If you can dedicate 2 (or more) channels to AF, particularly
with different excitation lasers (though the red laser on the calibur isn't great), you
can readily distinguish the AF+ cells since AF is typically excited over a broad range of
wavelengths and emission generally "correlates" on a bivariate plot to form a diagonal
"smear" in the 2 channels...though unfortunately you can't really use AF as a
"fluorescence parameter" in its own right, probably because it's a result of more than
one fluorochrome species with similar, but non identical spectral properties.

One last tip...when analysing your data it can be best to initially segregate cells on AF
level, if you are using a standard hierarchical gating	strategy instead of cluster
analysis etc, and only subsequently look at marker expression on cells with similar AF
levels.

Good luck!

Andrew

Andrew Mitchell Ph.D
Equipe Giovanna Chimini
ABCA1 transporters
Centre d’Immunology de Marseille-Luminy
CNRS-INSERM-Université de la Méditerranée
Campus de Luminy, Case 906
13002 Marseille, Cedex 9
France
Tel:  +33 (0) 4 91 26 94 90
Fax: +33 (0) 4 91 26 94 30
e-mail: mitchell@ciml.univ-mrs.fr
________________________________________
De : Eric Shaw [eseric47@googlemail.com]
Date d'envoi : mercredi 11 juin 2008 18:20
À : Cytometry Mailing List
Objet : autofluorescence and compensation

Hi Flowers,
I have a question.
I have a tissue dissagregation which will have lots of dead cells and various cell types
which most likely have different amounts of autofluorescence. I know I need to use single
stained controls (unstained, 7AAD, FITC,PE,APC) but wont my compensation be inaccurate
for FITC/PE/APC since I am compensating for the whole population- live and dead? And what
about the different cell types (epi's, tissue macrophages etc) that can give differing
amounts of autoflorescence? I compensate with my R1 gate  off FSC/SSC and I am using a
FACS Calibur. All replies will be appreciated.

Thanks,

Eric Shaw