[Cytometry] Compensation, Cy5PE, and APC
roederer at drmr.com
Fri Feb 7 07:47:13 EST 2014
I'd like to correct a few things here.
First, the Calibur was not "way ahead of its time when it came to time delay exclusion." This was already being done in the early 1970s with the first two laser machines in the Herzenberg lab, and all of the predecessors to the Calibur (FACS II, FACS IV, etc) did this as well.
Second, the inability to compensate between the parameters off different lasers on this machine was limited by the number of analog compensation controls offered on the instrument control panel, and the limitation on how much compensation (up to 100%) could be done. In reality, there was (and is) no issue with compensating the data after acquisition in software, and APC vs Cy5PE can be properly and fully compensated post-aacquisition.
What I refer to as "archaic" is not the technology, but rather than thinking that crossbeam compensation is something unique or problematic. It is no more problematic than single being compensation. (I recognize that many of the workhorse Caliburs are still in place today -- a testament to how robust and invaluable this technology has been).
If you believe that insufficient time delay resolution is a problem, then you would have to conclude that it would not be possible to use Cy7PE and Cy7APC in the same panel, because these are identical emission spectra in the far red region. and I should note that any cross beam "interference" will apply equally to the compensation controls in the sample: answer, this does not interfere with your ability to compensate; it simply reduces your sensitivity. This is why it's important to distinguish between compensation issues and instruments set up issues.
PS: I find it somewhat ironic that you take me to task for simplifying, whereas other people have taken me to task for the exact opposite (complicating their lives because they prefer to convey to their users that one cannot use Cy5PE and APC in the same panel, rather than to sit down and explain the consequences, and offer alternative solutions as options rather than requirements.) I stand by my statements; for the purposes of discussing Cy5PE vs APC in a panel, I don't think I over- or under-simplified.
On Feb 6, 2014, at 10:40 AM, Bill Eades wrote:
> Repeated valuable advice for the most part, Mario, but I would consider using a different reference for analog deficiency. In a world of controlled obsolescence, it may be proper to note that as archaic as the FACSCalibur may seem, it was way ahead of its time when it came to time delay exclusion. In the lion’s share of FACSCalibur placements, the instruments were in fact not optimized to the point where they could adequately compensate PE-Cy5 (Tricolor, CyChrome) from APC or Alexa 647. A fairly simple calibration for "pre pulse" could be performed to almost completely isolate the pulses emanating from delayed laser intercepts. I know for a fact that this calibration was routinely performed at NIH, and that the covering BD engineer as well as myself back in the day, made certain that this laser delay exclusion could work with PE-Cy5 and APC coexisting. Beyond that, there are a lot of FACSCaliburs still in use, or updated to digital, and that funding challenged cytometry sites will have to deal with being archaic. It is not such a bad thing, and, it is in fact, your roots. So, to your first point, there is a huge distinction between compensating parameters collected from different lasers than compensating primers collected on the same laser, if in fact, you have a FACSCalibur 4 color that has been properly set up by pre pulse adjustment. The starting values will be very different when taken to any hardware or software subtraction functions, and the resultant digital spreading will be less.
> In the furtherance of itemizing rant from reference, emission from cells is of course not collimated before it hits a laser intercept pinhole or fiber, which simply means that it can and will reach an adjacent intercept period, and its respective emission path. The amount of off axis cross beam interference has much to do with the instrument’s distance between intercepts, and in the “new world” of cramming intercept points within an objective lens, even more care has to be taken to decide what laser goes to what adjacent intercept. This is also the reason why dye choices should be nominally engineered not to incur direct spectral superimposition if at all possible. Mario, we know what you can do, but your absolute simplification will in fact be a stumbling block to the “newbies” of flow. Make the point, please, that is so informative regarding numbers versus graphic plotting, and that fear of compensation should be curbed, but do not forget that light is not nearly as manageable as a person would think, especially when we are leveraging accurate light detection with multiplexing demands. I have learned so much from you, which makes it so shocking to me that you tend to simplify so.
> William C. Eades
> Assistant Professor, Division of Oncology
> Washington University School of Medicine
> Director, Siteman Flow Cytometry Core Lab
> Office- 701B Southwest Tower
> Lab- 703 Southwest Tower
> Phone- 314.362.9364
> Shipping address:
> 703SWT, 4566 Scott Avenue
> Saint Louis, MO 63110
> On Feb 6, 2014, at 7:43 AM, Mario Roederer <roederer at drmr.com> wrote:
>> It's time for my biennial rant about the misunderstandings of compensation…
>> If you or someone in your facility thinks that Cy5PE and APC cannot be used in the same panel, then this is for you!
>> I was recently approached by a user who was told by his flow core that one could not use Cy5PE and APC at the same time because they have completely overlapping emission spectra, and it was not possible to do "crossbeam compensation". This thinking is archaic and arises from an inability of the old machines like the FACSCalibur to perform analog online compensation between FL3 (Cy5PE) and FL4 (APC). But we are in the modern world now, and all of our compensation is digital and software-controlled.
>> First: let's eliminate the term "crossbeam compensation" from the lexicon. There is no distinction between compensating parameters collected from different lasers then compensating primers collected on the same laser.
>> Second: In general, yes, you want to use dyes that overlap as little as possible. However, we have a limitation of number of dyes that are available and sometimes we don't have a choice. The more the dyes overlap, the more spillover-spreading you will have, and the less sensitivity you will be able to achieve after compensation. So don't use such combinations on antibodies that are expressed at low levels on cells because then you won't be able to resolve them from the background. This is part of how to build a multicolor panel; we have published many papers on this process and how to optimize it. But just because dyes overlap heavily doesn't mean they can't be incorporated into a panel.
>> Third: The reason that Cy5PE and APC can in fact be compensated despite the fact that the Cy5 and APC have nearly completely overlapping emission spectra, is because they have distinct excitation spectra. The amount of Cy5 emission appearing in the APC channel can be estimated from the Cy5PE measurement, and then be subtracted from the APC measurement, to leave only the signal coming from APC itself. This is what the compensation does. Fundamentally, in order to compensate to fluorescent molecules, all you need to have is either distinct excitation spectra OR emission spectra (or both).
>> Fourth: using Cy5PE and APC in the same panel is eminently doable. As long as you are aware of the impact of spillover spreading, you will be able to interpret the graphics just fine (but use numbers anyway). There are perhaps better choices of dyes: Instead of Cy5PE, use Cy5.5PE, Alexa680-PE, or Cy5.5PerCP if you can. These will have lower compensation values with APC than does CyPE.
>> (It might be noteworthy that we have used all of Cy5PE, Cy5.5PE, APC, and Cy5.5APC in the SAME panel -- with fine results. I'm not saying it's easy, but I am saying it's not impossible!)
>> Cytometry mailing list
>> Cytometry at lists.purdue.edu
>> Search the list archive at http://tinyurl.com/cytometry
More information about the Cytometry