I’ve been thinking about brightness in flow cytometry lately. It’s a simple enough concept, but it’s one of these things where you start thinking about it and you may quickly get lost in the numerous moving parts. One could argue that what people really care about is the level of separation between different clusters of cells. So while it’s important to maximize the brightness of the signal, it’s not the full story. There’s quite a few things to look into and I just decided to write down a list of the main factors impacting resolution. It’s a bit of a mess and I won’t explain much here, but you’ll find links if you want to dig a bit deeper.

The reagents

• One main variable of brightness is the fluorophore used. Some are quite dim, while others are very bright. This is a function of their chemical structure and a whole field within itself. Kelly Lundsten gives an overview of fluorophores families here. Another factor is the increased spreading error caused by the overlap between the emission of two fluorophores, which reduce the resolution of the clusters (see the very popular Trumpet Effect blogpost). Some fluorophores crush their neighbors, others gets trampled on: they are the perpetrators and the victims!
• The antibodies themselves may impact the brightness. Titration is your friend. Non-specific binding can increase the amount of noise from the negative cluster which is why blocking is so important. Improperly titrated antibodies can also be a a deterrent to good separation (or may cause you to overspend on your experiments!!)  Lastly, recombinant antibodies have been developed to limit the level of non-specific binding (see a Miltenyi example here).

The hardware

The instrument is another important factor when discussing brightness. It can be broken down in several parts:

• Lasers. the laser wavelength must reasonably match the excitation of the fluorophores to generate a good signal. Secondly, a laser with higher power will excite that fluorophore better (up to a point). And obviously, the laser needs to be properly aligned to maximize the illumination of each passing cell.
• Detectors. First off, you’ll have multiple types of detectors that will perform differently in portion of the spectrum (see Lawrence & al.)  Bert is working on a blogpost on the different detectors found on the instruments in the CAT Facility, so stay tuned. The voltage selection on those detectors will also play a role. Generally, increasing the voltage (or gain) on your detectors will push the positive and negative clusters apart. But at some point, the separation reach a zenith and increasing the voltage/gain any higher only moves the two clusters up the scale while simultaneously spreading the negative population. What you need to find is the sweet spot on the detector for your experiment where the separation is optimized. As we saw before, the spectral overlap between two fluorophores will decrease resolution, but proper voltage/gain selection may attenuate the issue. Finally, while detector position is fixed in most instruments nowadays, the detectors also have to be aligned to collect a maximum of light.
• Filters.  A fluorophore emitting at 530nm will not be picked up by a detector in front of which sits a 450/50, so one needs to keep an eye on these things. The width of the bandpass filters will also impact number of photons collected. So while an instrument with multiple detectors available on each laser line will allow to collect signals from a larger number of fluorophores, each filter may be narrower, thus limiting the collection of photons. And let’s throw in optical fibers in the flow cytometer. They can be used to direct the laser light to the flow cell, or to capture the emitted scattered light and send it to the detector array. The optical fiber will also need to be aligned, and it will deteriorate over time, resulting in lower signal transfer efficiency.
• Processing the signal. The way these photons are binned – that is, how the electronic board of the flow cytometer samples a signal and assigns it to a specific channel – is critical to resolution. A great review of this side of the instrument was published by Novo and Wood. More recently, Spectral Flow and the unmixing algorithm offers a new avenue to parse the signals coming from a sample. The algorithm, usually used on instruments optimized for these types of things, will allocate the photons coming from the different fluorophores using their signature – the aggregated intensity measurements of each fluorophore over the entire array of detectors in the instrument. It allows, for instance, separation of two closely emitting fluorophores that may differ in intensity in a few detectors only. While it’s still not a great idea to use these two fluorophores in the same panel, it’s enough to showcase how unmixing usually increase the resolution between population of cells.
• Sample focusing. Most flow cytometers use hydrodynamic focusing to align the sample in front of the laser beams. It’s a very elegant way of maximizing the illumination of the cells by the laser beams.  The pressure applied to the sample will impact the diameter of the sample core within the stream of sheath, so that a higher pressure lowers the ability to resolve small differences in intensity. One way to palliate this is to uses acoustic focusing, which is a feature available on the Attune NxT. Acoustic focusing uses sound waves to align particles before hydrodynamic focusing gets applied. So with the cells being already aligned, the pressure on the sample can be increased by quite a bit without loosing too much in the quality of the data. On the other end of this spectrum, our new MACSquant Tyto cell sorter focuses the cells in front of the laser beam by having them navigate a curving microfluidic channel, which is not as optimal. This not only increases the variance of illumination of each cell, but also the amount of noise generated. This is why it’s much harder on the Tyto to detect dim markers and resolve small difference in fluorescence intensity, and why we need to improve the brightness of the cells through other means.
• Let’s just assume that the instruments are working properly, shall we?

Biological Considerations

• Level of expression of the markers. For optimal resolution it is important to consider both the expression of the protein and brightness of the fluorophore when pairing markers and fluorophores . Protein expression can generally be characterized into three groups: primary markers are highly expressed and can be described as either “on” or “off”, secondary markers have a continuous range of expression, and tertiary markers are hardly recognizable from the background. Pairing a dim fluorophore with a tertiary marker will result in poor resolution.
• Co-expression. Can two very bright overlapping fluorophores be used together without any issues? It depends on marker co-expression and spreading error. The easiest scenario is to place highly overlapping fluorophores on markers expressed exclusively on two different types of cells. However with co-expressed markers, spreading error must also be factored into choosing the best fluorophores. Go see the Trumpet effect post again!
• Autofluorescence. A very high autofluorescence level can mask the brightest of fluorophores, and the use of the autofluorescence subtraction feature found on most spectral flow cytometers is likely to be critical in resolving the positive signal. Spectral Flow Cytometry is a great tool to get rid of the autofluorescence factor.
• We have a lot of training materials on this topic including: basic panel design, spectral panel design, and common panel design mistakes.

There’s a lot on the table and some of these topics gets complicated very quickly, which is why flow cytometry is so much fun. We’ll continue exploring these ideas here, so follow us on Twitter to see what is going on.