Re: Manual vs. Auto compensation

Hi Nidal,
we use to have frequent discussions about compensation issues at our 
institute (as probably all flow based facilities do ;) ) and I will just 
pass on some thoughts I picked up during my time here. First of all, 
from personal experience, autocomp works just fine. I just recently 
compared the two (manual and auto) when teaching one of our younger 
technicians how to use our FACS Canto and we therefore did it the quick 
way and the 'proper' way and guess what.....we ended up with almost the 
same compensation values. So, either way, if you do it properly, it 
should work. The manual guys are normally the ones having started flow 
cytometry in the old FACSScan or Calibur days and refuse to 'let go', 
but the thing is, even when you mess up your compensation on the new, 
digital machines, you can still work on that AFTER your measurement. 
Much more important is the setting of the PMT-voltages, this is the only 
thing you have to make sure that your signal to noise ratio is ok with 
your settings. Some people at our lab don't even bother compensating the 
samples while measuring and just take the single stains over to their 
FlowJo analysis and do the compensation there. That's sort of 
autocompensation, too, since it's the same matrix based system as with 
the DIVA software (I don't know about the underlying algorithms, 
though). To see whether you have over or undercompensated your sample, 
you could use the biexponential display, which gives you an idea whether 
you have produced 'negative' cells when overcompensating the signals. 
Just makes plots prettier though, since your cells won't all sit on the 
axis, but give you a nice cloud instead. The thing is, that plots may 
look different when you change your compensation procedures (and since 
people just agree to what they are used to, different may be a 'bad' 
thing ;)  ), but as long as you can make out your population, it should 
be fine. Only problem of course would be if you look at expression 
strength of markers, since it might be difficult then to find out what 
is high expression, low expression or no expression at all, that's 
something you would have to establish for the specific scientific 
question you want to answer.
In terms of using different cells for compensation, you are right when 
it comes to the issue of autofluorescence....BUT....the difficulty may 
be, that in your sample you just don't know if your specific markers 
really are detectable and (if not) may leave you with no clear positive 
staining whatsoever or cell fractions so small that you just can not use 
them for compensation. We had the discussion whether it is ok to switch 
to a different marker then labeled with the same colour (eg using CD8 
FITC instead of CD86 FITC) and this should be ok at least with the 
non-tandem dyes (I will come to that issue in a second). All you have to 
make sure, that you could detect both FITC-signals with the same 
PMT-Setting, the spillover should be the same then and you get a nice 
and clean positive signal when using a strong marker. I would rather (as 
you stated) use my tissue sample and switch my Abs rather than sticking 
with my markers I use for the analysis and take a different tissue (eg 
spleen instead of liver). That rules out the autofluorescence issue. 
Only problem with switching markers is the use of tandem dyes, simply 
because the FRET-efficiency depends on the quality of the conjugation 
and the amount of dye degradation (which is unavoidable after some 
time). Old tandem conjugates can be particularly nasty when using them 
together with the single dye (eg PE together with PE-Cy7). The 
degradation of the PE-Cy7 would leave you with a perfect PE-signal and 
that would be impossible to discreminate from your PE-labled Ab. So it's 
quite useful to check your tandem dyes from time to time (and keep them 
dark and cold as much as possible).
Hope that helps, that's as far as I've gotten through the jungle of the 
flow-cyto-how-to's

regards
Felix


Felix Heymann
Institute for molecular medicine and experimental immunology
University clinic Bonn, Germany



> Hope all are having a great weekend and not in the lab on a Saturday 
> night.
>
> I'm a relatively knew guy to the sort/flow world and need to set some 
> things straight so I hope those with more experience can give me their 
> feedback.
>
> I have people in the lab who complain about me not doing manual 
> compensations and instead have the Aria calculate them for me. And 
> when I have the Aria compensate, they want to manually compensate 
> after they record the data because the plots "don't look right." I 
> mean, isn't that simply producing "make-belief" data? Or am I missing 
> something? I'm not trying to criticize those who do manual comp. I 
> just want to understand why do I need to manually compensate when it's 
> all based on mathematics and can be done by the machine?
>
> And the frustrating part is that when they bring me the compensation 
> samples, they contain cells that are different from the ones in their 
> analysis samples. I try to tell them that the autofluorescence of both 
> cell types is most probably different and that might be the reason why 
> some of the compensations don't turn out correctly. But they dismiss 
> my suspicion, claiming that it should not affect the compensation 
> since we're dealing with colors and not cell density. Is that true?
>
> I'm stressed out. I need some ice cream.
>
> -- 
> Nidal Muvarak
> Abramson Research Center
> Children's Hospital of Philadelphia
> 3615 Civic Center Blvd
> Philadelphia, PA 19104