Written By Tim Bushnell, Ph.D.

Good flow cytometry depends on a high-quality, single cell suspension. If the cells put through the instrument are not of high quality, the ensuing data will be difficult to analyze.

Likewise, if the sample is clumpy, one will not be able to readily distinguish cells of interest from the clumps they are attached to. Thus, sample preparation becomes the critical first step in any flow cytometry experiment.

As Henry Ford once said, “Before everything else, getting ready is the secret of success.”

For the purposes of this article, our attention will be focused on mammalian cells and cell-lines. While other samples can benefit from these tips, samples from the environment, bacterial and yeast samples, and small particles like extracellular vesicles, have their own issues and tricks to consider.

To get high quality results, follow these 3 sample preparation steps…

1. Get the purest sample possible.

Taking the example of working with peripheral blood, one of the first and most important choices to be made when using peripheral blood is the type of collection tube that will be used.

Since most of the tubes used for human blood have some form of anticoagulant in them, the choice of tube also dictates the anticoagulant. You can choose EDTA or heparin as the anticoagulant, although each researcher may have a personal choice for anticoagulant based on downstream applications.

In this example, the sample contains red blood cells. If you study RBCs, this is great. If not, the next step is some removal process.

The easiest method is to use some form of lysis solution — for example — ACK buffer. This lyses cells based on osmotic differences between RBCs and leukocytes.

While you can buy this solution from many vendors, there are a lot of recipes on the Internet. Here is one ACK recipe to try…

The protocol to use this is to add 1 ml of 1X ACK buffer and suspend the cells well. Wait 5 minutes and add 10 mls of Flow staining buffer (FSB: 1xPBS with 0.1% BSA), invert a few times and centrifuge.

This is especially good if working with other RBC containing samples like bone marrow or spleen. The timing of the protocol is critical, as is testing the reagent vigorously to ensure it is not lysing your cells of interest.

Figure 1: CPT tube separation demonstrating the different layers.

Where Ficoll may have been popular in the past, there is movement to favor having samples collected into the BD CPT tubes. These are great tubes to use because they combine a density gradient with a special gel.

When the tubes are centrifuged, a density separation of the cells occurs, and the gel barrier separates the RBCs and neutrophils from the Lymphocytes and Monocytes — making removal very easy.

Boyd et al (2015), in a paper, discusses different ways to improve detection of Salmonella in blood, and shows a great picture in figure 2 that illustrates the results of different lysis/isolation methods.

2. Count your cells correctly.

Once the single cell suspension is prepared, it is time for the dreaded counting of the cells. This is an essential step, as the recovery may make or break the experiment. It is also good to count the cells before staining and after staining to see what losses have occurred. If you’re not careful, you can see losses of up to 30% per centrifugation step.

These counts are critical to make sure that there are enough cells for the downstream applications. This is especially true if the experiment is sorting, and a specific minimum number of cells is needed.

Three methods for counting cells include:

• manual (using a hemacytometer — the ‘gold standard’),
• image-based (instrument names), or
• flow-based (counting beads or pump-based system).

Each method has its strengths and weaknesses, so should be fully qualified and consistently applied. Consider using both image-based (using the Countess) and flow-based (Accuri).

As with every method, one must qualify the assay, typically using either a hemacytometer or other counting standard. If the system is kept in good condition, results like these, provided by Claire Rodgers (when she was at Accuri), are possible with Accuri.

Figure 2: Cell counting comparison using counting beads and the Accuri C6.

One other important advantage of using a cytometry-based method is that dead cell enumeration is easier with a cell-impermeant dye (PI, for example) rather than relying on Trypan Blue.

Figure 3: Structure of Trypan Blue

While this dye has been used for many years, the issue becomes determining how ‘blue’ is a dead cell. Of course, having RBCs in the mix can also be an issue, as shown below.

Figure 4: PMBCs stained with Trypan Blue with (left) or without (right) red blood cell ACK lysis.

3. Prevent your cells from clumping.

The final step in having high quality single preparation is to take steps to ensure the cells don’t clump. There are several common causes of this and even easier fixes.

1. Dead cells — Dead cells can release DNA into the medium. DNA is the duct-tape of the biological world and will easily cause clumping by binding cells together. The easiest solution for this problem, and one recommended especially for cell sorting, is to add 10 units of DNAase per ml of sample to minimize this effect.
2. Cations — Calcium and Magnesum can promote cell clumping. When at all possible, use Ca++ and Mg++ free PBS for your staining buffer, and consider adding 1 mM EDTA to your buffers as well. Not only will this help reduce clumping, if you’re going to be fixing cells stained with QDots, it may be essential, as Zarkowsky et al., showed in 2011.
3. Over pelleted cells — If cells are pelleted too hard, they can clump. Make sure to know the relative centrifugal force of each centrifuge the samples are prepared on and to use that as a guide for setting the speed of the centrifuge. Remember different rotors will have different RCF.
4. Clumped cell removal — At the end of the day, if there are clumps of cells in the sample, they need to be removed. Filtration is the fastest and easiest way to do that. Mesh strainers are available from several vendors, but can be very pricy. You can consider ordering in bulk from a company called ‘Small Parts’ on Amazon where you can get mesh like this 50 micron sheet for a reasonable price. Cut it up into convenient sizes and filter as needed. Prewet the mesh and put the pipet tip close to the filter to ensure the sample went through.

In the end, keeping these three steps in mind will improve the final outcome of the experiment — by ensuring that the cells are in a single cell suspension, and clumps don’t clog the instrument. Remember Howard Shapiro’s 1st Law of Flow Cytometry (from Practical Flow Cytometry, 4th Edition, page 11)…

In the end, good flow cytometry depends on a high quality, single cell suspension. If the cells put through the instrument are not of high quality, the ensuing data will be difficult to analyze. Likewise, if the sample is clumpy, one will not be able to readily distinguish cells of interest from the clumps they are attached to. Sample preparation becomes the critical first step in any flow cytometry experiment.

To learn more about flow cytometry protocols and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D.

Do you annotate your data properly?

If you don’t, you’re not alone. Most people don’t take advantage of the keywords embedded in the FCS (Flow Cytometry Standard) files to annotate their data. FlowJo makes this easy to do.

First off, what are keywords and why do you want to use keywords?

The flow cytometry standard file formats (FCS) was developed by the International Society for the Advancement of Cytometry (ISAC) to create a standard file format for all cytometers to use. The purpose of this was so that no matter who manufactured the instrument, or the analysis software, files could be created and read in any system.

As part of this FCS standard, ISAC created a set of core keywords that MUST be used. These include information on the parameters, scaling, FCS file format, date, file name, and numerous others. You can read more about the FCS file format here.

When it comes to annotating your flow cytometry data with keywords, there are 3 things you need to know…

1. Certain keywords are “required” by the Flow Cytometry Standard (FCS).

This is the list of primary keywords required:

$BEGINANALYSIS Byte-offset to the beginning of the ANALYSIS segment.
$BEGINDATA Byte-offset to the beginning of the DATA segment.
$BEGINSTEXT Byte-offset to the beginning of a supplemental TEXT segment.
$BYTEORD Byte order for data acquisition computer.
$DATATYPE Type of data in DATA segment (ASCII, integer, floating point).
$ENDANALYSIS Byte-offset to the last byte of the ANALYSIS segment.
$ENDDATA Byte-offset to the last byte of the DATA segment.
$ENDSTEXT Byte-offset to the last byte of a supplemental TEXT segment.
$MODE Data mode (list mode – preferred, histogram – deprecated).
$NEXTDATA Byte offset to next data set in the file.
$PAR Number of parameters in an event.
$PnB Number of bits reserved for parameter number n.
$PnE Amplification type for parameter n.
$PnN Short name for parameter n.
$PnR Range for parameter number n.
$TOT Total number of events in the data set.

Many of these keywords have little value to you, as the user. They can help with troubleshooting, if you are having trouble in the analysis program, but for the most part, not many people need these keywords.

2. Other keywords are optional and can assist you in analyzing your data.

The keywords that will help you analyze your data are the optional keywords. Although these keywords are considered optional, most cytometer manufacturers still use them.

$ABRT Events lost due to data acquisition electronic coincidence.
$BTIM Clock time at beginning of data acquisition.
$CELLS Description of objects measured.
$COM Comment.
$CSMODE Cell subset mode, number of subsets to which an object may belong.
$CSVBITS Number of bits used to encode a cell subset identifier.
$CSVnFLAG The bit set as a flag for subset n.
$CYT Type of flow cytometer.
$CYTSN Flow cytometer serial number.
$DATE Date of data set acquisition.
$ETIM Clock time at end of data acquisition. ISAC Recommendation FCS 3.1 – Data File Standards for Flow Cytometry 11/34
$EXP Name of investigator initiating the experiment.
$FIL Name of the data file containing the data set.
$GATE Number of gating parameters.
$GATING Specifies region combinations used for gating.
$INST Institution at which data was acquired.
$LAST_MODIFIED Timestamp of the last modification of the data set.
$LAST_MODIFIER Name of the person performing last modification of a data set.
$LOST Number of events lost due to computer busy.
$OP Name of flow cytometry operator.
$ORIGINALITY Information whether the FCS data set has been modified (any part of it) or is original as acquired by the instrument.
$PLATEID Plate identifier.
$PLATENAME Plate name.
$PnCALIBRATION Conversion of parameter values to any well defined units, e.g. MESF.
$SMNO Specimen (e.g. tube) label.
$SPILLOVER Fluorescence spillover matrix.
$SRC Source of the specimen (patient name, cell types)
$SYS Type of computer and its operating system.
$TIMESTEP Time step for time parameter.
$TR Trigger parameter and its threshold.
$VOL Volume of sample run during data acquisition.
$WELLID Well identifier.

Many of these keywords have values that are recorded by the cytometer. However, many of these keywords are also empty, so you can put in whatever you want during acquisition! PatientID, treatments, stains, studyID, groups, etc. Think about what may help you in your analysis, or to track your data when you export it. You should consider putting in these keywords.

3. You can create keyword formulas to simplify the data analysis process.

To begin working with keywords in FlowJo…

• Load some data in FlowJo and in the workspace
• Right-click on a sample
• Go to Sample Info (or Inspect). An interface will open that displays all the keywords and their values.

• Keywords like $BTIM and $ETIM denote start time and end time.
• $DATE is self-explanatory.
• $FIL is the file name.
• $TOT is the total number of events that were acquired.

These are just examples of commonly used keywords. FlowJo has a full list of keywords on their website, here.

Keywords are embedded into the FCS files, so you can’t edit them. However, FlowJo does allow you to add them as a column in the workspace and edit them for that workspace.

How can this benefit you?

If you forget to name a stain properly, add the $PnS keyword and put in the name. This can be a little challenging in that you will need to know the channels. These can be found in the sample information menu under $PnN usually.

The $PnN is the parameter name, but isn’t often very useful, as this is a default name set up on the cytometer. It may just be FL1, or maybe the filter set, like 525/50.

If you add a value to the corresponding $PnS keyword though, that will show up in the graph window’s axis labels and make it easier to review the data.

You can also add new keywords in FlowJo. Just click the add new keyword button.

You can use this to add sample IDs, or patient numbers, treatment information, staining protocols, or other information needed for you analysis.

You can then group samples based on keywords. If you’ve never used groups, that is also something that allows you to batch process your analysis to a subset of samples.

Finally, you can embed keywords into the layout editor and the table editor in FlowJo. Use the table editor to add a formula of $ETIM-$BTIM to get the time of acquisition.

1. First, add $ETIM and $BTIM to your table.

2. Then go to Add Column > Formula.

3. Insert reference of $ETIM – $BTIM and click OK.

The formula will appear in the table editor and when the table is created, you’ll get the column in the table. Use $TOT and divide by $VOL to get the count per unit volume. Embed $SAMPLEID to be able to easily sort samples when you put them into Excel from the table editor.

There are numerous different ways to use keywords in the table editor and layout editor. The problem is, most scientists do not annotate their data properly and pay the price when they want to repeat their experiments. By taking advantage of the above keywords and by using keyword formulas, you can save time during your analysis. Most importantly, when you go to reanalyze your data, you can utilize your previous keywords and formulas to save even more time.

To learn more about how to annotate your data with FlowJo keywords and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D.

In the most common applications, cell sorters are utilized to purify cells on the order of about 5-30 microns in diameter, often human or murine leukocytes, or cell lines derived from liquid or solid tissue types.

Considering this typical size range, cells between 5-10 microns in diameter are typically the simplest cells to sort. That being said, the general rules of thumb in forming sample preparation, instrument setup, and sample collection are critical to generate sorts of high purity, recovery, and viability — the hallmarks of sort quality.

To this end, there are 4 steps you must take to ensure high quality when sorting 5-10 micron cells by flow cytometry…

1. Protect cell sample quality.

The tried-and-true adage of flow cytometry — garbage in, garbage out — should hover above and firmly guide each step of the sample preparation process for any kind of cytometry experiment, but is especially critical for sorting.

Poor sample quality will most surely guarantee sacrifices in purity, recovery, and/or viability, often to the extent of significantly, affecting the downstream experiment, so follow these guidelines for the best outcomes.

The key is to minimize debris because the amount of “junk” in a sample can often severely limit the purity, efficiency, and even viability of a sort.

As a reminder, the efficiency parameter reported by the instrument indicates how many target cells (cells in the sort gate) were sorted compared to how many target cells were thrown away in order to maintain high purity of the output.

The efficiency is highly dependent on event rate: the higher the event rate, the higher the chance that a non-target cell will fall close enough to a target cell that it may be sorted along with the target cell. These types of events are aborted (not sorted) in purity sort modes.

Even though debris particles often do not fall into the sort or even scatter gates, they do register as events. Since these events are not classified as sort events, any debris particle in the same droplet or close to a target cell’s droplet will result in an abort or coincidence of that target cell.

Debris can be minimized in the following two simple ways:

1. Keep cells alive during preparation.

Dying cells will lyse and fracture into debris. Keep cells cold during preparation, assuming the cell type permits. Avoid rough handling, especially when preparing a suspension of adherent cells.

Avoid scraping and expose cells to proteolytic enzymes for dissociation for the minimal amount of time necessary to remove cells from the growth surface or to disaggregate solid tissue.

1. Use a gradient or kit to remove debris if necessary.

Preparing cell suspensions from solid tissue can often generate debris. Centrifugation gradients or specialized debris removal kits can be a boon in removing unwanted particles from the sample.

Miltenyi Biotec offers a Dead Cell Removal Kit, for example, that works quite well. Disaggregated neural tissue can generate a lot of debris, and there are resources from Miltenyi and Worthington that are helpful.

2. Optimize flow cytometry instrument setup for cell type.

Instrument setup is not especially unique for cells in the 5-10 micron range compared to larger cells, so the usual guidelines for optimizing the instrument apply, including choosing the most appropriate nozzle.

Nozzles come in several different size ranges on cell sorters: 70, 85, 100, and 120-130 microns. The smaller the nozzle, the higher the pressure that is required to generate stable droplet breakoff the necessary distance from the nozzle orifice.

Larger or fragile cells may require lower pressures and larger nozzles to maintain purity, recovery, and viability of the sorted fraction, but many cell types in the 5-10 micrometer range can be sorted using the 70 micron nozzle and the higher pressure ranges required (~70 PSI).

The general rule that cytometrists follow when selecting a nozzle is: “the nozzle size should be no smaller than 5 times the diameter of the cell.”

Adhering to this rule, the 70 micron nozzle would be appropriate for cells up to 12-15 microns.

One nice benefit of the 70 micron nozzle is sort speed. The smaller the nozzle size, the faster that droplets can be generated.

Droplets effectively partition the stream into “sort units,” so the faster they are generated, the faster we can sort. The frequency of the droplet drive, which determines how many droplets are generated per second, can be as high as 90 kHz (90,000 droplets per second) at 70 PSI and can permit sample rates of ~30,000/second for sorting.

That being said, 70 PSI may be too high for some cell types in the 5-10 micron range. Try a larger nozzle and lower pressure if consistently poor viability is observed in the sorted fraction.

3. Always gate for single cells.

When setting up the sort, always include a doublet discrimination gate in the sort logic.

Doublets are events in which two particles pass through the laser spot so close to each other that they are classified by the instrument as a single particle and can contribute to low purity.

A doublet may consist of a target cell and a non-target cell, both of which will be sorted.
Because the instrument classifies both cells as one event, the presence of the non-target cell will not result in an abort.

A typical doublet discrimination gate plots the area pulse parameter against the height or width pulse parameter in a single channel, often forward scatter or side scatter (e.g. FSC-A x FSC-H). This allows the signal intensity to be discriminated from the time the particle spends in the laser.

Doublets often generate the same intensity in FSC and SSC as single particles but spend close to twice as long in the laser spot. It can be helpful to create two doublet discrimination plots, and gate hierarchically (SSC-A x SSC-H gated on FSC-A x FSH-H).

4. Keep cells alive and healthy during and after the sort.

Follow the simple steps below to ensure high purity, recovery, and viability of the sorted fraction.

1. Suspend the pre-sort sample in the right sample buffer for sorting.

Choose a suspension liquid that is most appropriate for your cell type, but avoid those containing CO₂-carbonate buffering systems, which are typical components of cell culture media.

These buffering systems are formulated for tissue culture CO₂ partial pressures and not atmospheric, so they can become alkaline over time and may contribute to cell death.

Media containing phenol red will turn purple under these conditions, indicating alkalinity.

PBS and HBSS, without calcium or magnesium, are good choices, and be sure to include some kind of protein, typically 1-2% FBS or BSA. Calcium and magnesium promote cell adhesion and stickiness, which is to be avoided at all costs. Protein helps keep cells viable and may prevent sticking as well.

1. Use a viability dye, always.

A viability dye like propidium iodide, DAPI, 7-AAD, or many of the other commercial or proprietary dyes, will not only help assess the cell health of the pre-sort sample but will also allow you to prevent sorting dead or dying cells along with viable cells.

Dead cells will shed markers or leak GFP, so they may appear non-fluorescent after the sort even though they appeared in the sort gate in the pre-sort sample, leading to an assessment of low purity. Keep in mind that the scatter plot is not always indicative of viability.

Cells that are dying may appear in the same population as the viable cells, while dead cells that have fractured will appear outside of the scatter gate configured for viable cells.

1. Never collect sorted cells in an empty tube.

Cells are traveling at many meters per second as they are sorted, and if they hit plastic at these velocities, they will be obliterated. Including a collection buffer in the collectIon tube will drastically improve post-sort viability.

The buffer can be the same buffer that is used to resuspend cells, but be aware that cells can be accompanied by significant volume, which can dilute the collection buffer.

If large numbers of cells are to be collected, one strategy is to use PBS or HBSS containing a high percentage of serum, considering that it will be diluted to normal concentrations once the collection tube is filled.

While sorting cells 5-10 microns in diameter does not present a particular challenge compared to other cell types, the standard procedures must be followed to ensure the hallmarks of a good sort — purity, recovery, and efficiency. These good practices, including protecting sample quality, optimizing instrument setup, gating on single cells, and always keeping your cells happy will guarantee quality sorts, time and time again.

To learn more about flow cytometry procedure for accurate sorting of 5-10 micron cells and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

Flow cytometry is a powerful technique impacting both clinical and research.

If you’re looking for a career, flow cytometry technology can take you many places. An experienced flow cytometrist can find a job in a biotechnology company, in academia, or in a clinical setting.

Rather than focusing on specific career information, I would like to highlight one career option that has been very good for me since my transition from research scientist to core manager over a decade ago.

Core facilities, or Shared Resource Laboratories (SRL) as this Cytometry A paper made popular, represent an investment by institutions in resources and personnel.

Staff and directors working in SRLs have the advanced training and experience to support the researchers and research mission of the institution.

Working in an SRL is a different and very exciting career option for researchers who enjoy flow cytometry, enjoy working on many different projects, and enjoy working in scientific customer service.

That being said, starting out in an SRL can be daunting. Professionals of all kinds, or “users”, expect you to be an expert and will come to you for answers to all sorts of questions.

Some of these questions are easy to answer, some are not. As a result, a key to being successful in a flow cytometry career comes down to staying up-to-date on the latest information, best practices, top resources, and more.

When it comes to starting a career in flow cytometry, there are 5 strategies that will help you hit the ground running…

1. Learn all you can about flow cytometry technology.

Education is critical for making the most of any scientific technology. In flow cytometry, learning everything from how the instrument works, to gauging fluorochrome brightness, to troubleshooting problems with the instrument, with experiments, and with data analysis.

There are a large number of ways to get this education, some of which include:

• Vendor Training Courses — these are great for instrument specific training.
• Annual Training Courses — there are several courses around the world that are put on by regional societies. These courses are a great way to meet experts in the field, network with peers, and learn new techniques.
• Society Training — ISAC, CCS, and ESCCA have various training materials that are accessible to researchers. Check out the links to the societies above to learn more.

2. Get involved in regional users groups.

Finding all the local users of the technology can be challenging. Many areas have a regional cytometry group that hosts an annual meeting. This is where you can find local users and experts of the technology. If there isn’t one, consider starting one.

Meetings like the CYTO and ICCS annual meetings are good for networking while learning the latest and greatest in the field of flow cytometry.

3. Build your support network.

There will be questions and situations that will arise out of your realm of experience. A support network where you can ask questions and get answers in a timely fashion is critical, especially in the early stages of a career in an SRL.

There are a few very good resources that should be at the top of the list. These include:

• The Purdue Listserv — Over 20 years and still going strong. This is the ‘List’ for users of cytometry to ask questions and get answers. With a searchable database, it is an excellent place to learn about topics from sheath fluid, to isotype controls, and everything in between. The list is moderated, allowing for more focused conversation, while keeping things on topic.
• The Cytometry Google group — Started a few years ago, this group offers an alternative to the Purdue list. This group is growing and has a different feel than the Purdue list. The ease of uploading pictures, allowing vendor responses (and advertisements), and the power of Google (such as video ‘hangouts’) all offer another complementary way to get information.
• The Expert Cytometry Mastery Class — The Mastery Class is the world’s fastest growing and most successful flow cytometry training program. In addition to the Mastery Class offering an annual subscription to a 4-module training course and a live webinar series, members can get access to a private discussion group where on-going training and education topics are covered, where questions are answered in a matter of minutes, and where strong professional connection are made.

4. Identify a mentor and ask for support.

If you are new to the field, finding a mentor is a very useful step. Your mentor should be a senior scientist who can help you navigate the greater process of being a purveyor of technology.

This can include career advice, someone to turn to for help when solving difficult scientific or technical problems, someone to bounce ideas off of, and more. If they are at your institution, they can help with potential ‘political’ issues that will arise. As you move up the career ladder and become the manager or director of a facility, this support is very useful.

5. Travel to expand your network and expertise.

It costs money, it takes time, and you may have to jump through hoops to get approved at your institution, but it’s difficult to be a master of all techniques where you are, especially if resources or expertise are limited, and you may need to travel to learn what you need.

Sometimes, the better strategy is to go to another institution to learn a specific technique, rather than beat one’s head against the instrument trying to get things to work.

Becoming a member of an SRL is an exciting opportunity. There are many different ways to get into this field. To be successful in the field, seek out new educational opportunities and network with your peers. Flow cytometrists in the field are more than willing to talk and share their ideas and experiences while helping out a fellow cytometrist. A little effort can go a long way.

To learn more about how to get a flow cytometry job and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

Fluorescence compensation is one of the more difficult, understandably confusing, and misunderstood aspects of flow cytometry.

Compensation is almost always necessary when more than four or five colors are measured simultaneously. As such, understanding what compensation is, why it is necessary, and what to expect when using it, are critical for generating useful and high-quality data from flow cytometry experiments.

In order to understand compensation correctly, you need to understand the answers to 3 core questions related to compensation…

1. What is the nature of fluorophores?

The necessity of compensation arises from the nature of fluorophores, the fundamental and cherished tools of flow cytometry.

We often tend to associate a fluorophore with a particular color. For example, we often say that “FITC is ‘green’” and “APC is ‘red.’” While it is true that when FITC is excited it is most likely to emit photons in the “green” range (around 520 nm), it is also true that FITC will emit photons of other wavelengths of colors, from about 480-600 nm. However, the number of “non-green” photons that FITC emits gets lower and lower the farther the wavelength is from the emission maximum of FITC of ~520 nm.

The “spectral promiscuity” of fluorophores is a problem for flow cytometric measurements. A spectral plot, which describes the excitation and emission spectra of a fluorophore, is a powerful tool to understand why.

The plot shown below, reproduced from the BioLegend Spectra Analyzer, illustrates the emission spectra of two commonly used 405 nm excitable fluorophores: Brilliant Violet 510™ and Brilliant Violet 605™. Additionally, the spectral plot also illustrates the wavelength ranges (spectral bands) of bandpass filters that can be used to measure these fluorophores.

Notice how wide the curves are for both fluorophores and especially how wide the emission profile is for Brilliant Violet 510™.

This fluorophore, although mainly “green,” emits photons of all sorts of colors — from <450 nm all the way up to around 700 nm. That is quite a range! This wide span of emission wavelength quickly becomes a big problem as more and more colors are used.

The spectral detection range used in flow cytometry is about 370 nm up to about 800 nm, and there are only so many ways that this distance can be partitioned into spectral regions.

Because of this, spectral overlap, or the presence of one fluorophore in another’s detector, is inevitable in multicolor cytometry.

Specifically, in an experiment where two fluorophores are used, A and B, and two detectors designated to measure those fluorophores, D_A and D_B are used, A will be measured not only in D_A but a portion of A’s fluorescence will also be measured in D_B. Likewise, depending on the filter and the molecule, B may be measured in both D_B and D_A.

As illustrated in the spectral plot above, because of the wide emission profile of Brilliant Violet 510™, it will be detected in a bandpass filter about 85 nm away.

2. What is the spillover coefficient?

This reality of fluorophores would render data from multicolor experiments useless if something weren’t done to deal with it. That thing is compensation, which is a mathematical operation that accounts for and subtracts the spillover signal (i.e. the signal from other fluorophores) in every detector used in any flow cytometry experiment.

This process can be performed either manually or automatically, the automatic route being more preferable and accurate, especially as more and more colors are used.

While the comprehensive mathematical basis of compensation is beyond the scope of this article, taking the time to understand how it works from a high-level leads us one big step away from the “black box” model of interacting with a cytometer.

As always, the more transparent the box, the better the cytometry and the better the science.

The basic principle of compensation relies on one critical concept…

The amount of spillover signal in a secondary detector is always proportional to the amount of signal in the primary detector, regardless of the amount of fluorophore present.

In other words, in a two-color experiment utilizing fluorophore A and fluorophore B, if 20% of A’s emission spectrum is measured in B’s detector, we can always calculate the true signal of B in its detector, no matter how much of each dye is present, by co-opting the constant proportionality between A’s signal in B’s detector and A’s signal in A’s detector; the signal in from A in B’s detector is always 20% of A’s signal in its own detector.

To calculate this proportionality coefficient, properly called the spillover coefficient, we use single-color controls that contain only one fluorophore. By doing this, we can then measure the signal from this fluorophore, independently of any other fluorophore, in all detectors which facilitates calculation of the spillover coefficient.

Let’s look at the math to parse this out a bit. Stick with us here, even if math stirs up some trepidation.

The explanation below is from a report in a Current Protocols of Cytometry and it illustrates two-way compensation between two detectors, and assumes that we are using two fluorophores: fluorophore A and fluorophore B. While this is an atypically simple explanation, the principles described below carry through to the broader method compensation of any number of parameters.

We will use a few abbreviations, defined as follows:

^AF_A = signal from fluorophore A in its own detector (detector A)
^AF_B= signal from fluorophore A in B’s detector
^BF_B= signal from B in its own detector
^BF_A= signal from B in A’s detector
D_A = total measured signal in A’s primary detector
D_B = total measured signal in B’s primary detector

Therefore, the measured signals in each detector are:

The spillover coefficients for A’s and B’s detectors, S, can be defined as follows:

In order to actually calculate the spillover coefficients, we need to measure signal from each fluorophore independently and thus need single color controls. The total signals, from a control containing only fluorophore A, in each detector is:

Therefore, for A’s single-stained control:

Similarly, for B’s single-stained control:

Once the proportionality constants are defined, the “real” fluorescence in any detector in a fully-stained sample can be easily calculated by rearranging equations. Easy!

In the fully-stained sample:

All we need to do now is to do some crafty algebra to obtain the actual amount of A (^AF_A) in detector A and the actual amount of B (^BF_B) in detector B.

3. What is a compensation matrix?

When only two colors are used, the math is simple — it’s just algebra.

However, there is a way of calculating compensation for any number of parameters, which applies to more real-life cytometry experiments when many colors and detectors are used.

This method is an extension of the equations described above and uses linear algebra to generate a spillover coefficient matrix for all parameters being used to measure fluorescence in the experiment.

The math used in this case requires that the spillover coefficient matrix be inverted, which ends up as the compensation matrix.

The difference between the spillover coefficient versus the compensation matrix have been described in the same Current Protocols of Cytometry report:

Because the compensation matrix is more relevant to cytometry data — it tells us what to subtract from the detector in order to see the true fluorescence from its primary fluorophore — it is what is displayed in cytometry acquisition and analysis software.

You won’t ever have to perform these calculations yourself as, whether you are compensating manually or automatically, all of these calculations are performed by the computer. That being said, the cytometrist’s job is far from complete.

The basics of compensation as defined here provide foundational understanding of flow cytometry compensation. What it is, why it’s necessary, and what to expect make up the introduction to this series. Stay tuned for the next article on compensation for a deeper look into controls, the nuances of compensated data, and some mythbusting.

To learn more about why flow cytometry fluorescence compensation is critical for quality measurements and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

Flow cytometers have three major components:

• Fluidics that move the cells from the sample tube to the intercept point
• Optics that collect the light
• Electronics that convert the photocurrent into a digital value that is stored for later analysis

If we look at the fluidics system of a standard flow cytometer, laminar flow is established by the sheath fluid. This is even flow with fluid flowing in ‘parallel layers’. Due to some friction on the side of the tubing, the flow ultimately develops a parabolic shape, as shown below.

Into the center of this laminar flow, the cells are injected. The cell and fluid mixture is introduced at a higher differential pressure, keeping the cells in the center of the laminar flow, and allowing the process of hydrodynamic focusing to occur, which causes the cells to spread out along the velocity axis, single file, as they approach the intercept point.

After the intercept, the cells either flow into the waste or, in the case of a cell sorter, the stream is broken into droplets, and the appropriate droplet is charged and sorted.

Most of the interactions that a user has with a flow cytometer is with the fluidics system, and many of the issues that users will face in troubleshooting problems on the instrument will also be here.

Here are four important questions to ask yourself when trying to understand and troubleshoot the fluidics system in your flow cytometer…

1. What sheath fluid are you using in your cytometry protocol?

Many institutes run a phosphate buffered saline (PBS) as their sheath fluid. This can be made in-house or purchased from any number of vendors. However, since the sheath fluid and sample stream do not mix, it’s not necessary to use PBS for the sheath fluid.

Others use just water for the sheath. Several years ago, we converted to using water with 0.1% 2-phenoxyethanol, which is used as a preservative and has some surfactant properties.

For cell sorters, however, one must use some buffered saline solution for the sheath fluid. Since 2004, I have used a 10 mM HEPES buffered saline solution for my sorting needs. This is because HEPES is a better buffer at sorting pressures than Phosphate.

If your cells are prepared and held in a culture media (like RPMI), adding HEPES to the solution is appropriate, as culture medias are typically formulated to buffer in a CO₂ environment.

Our solution (10X) is as follows:

• 13.015g HEPES sodium salt
• 11.915g HEPES free acid
• 80g NaCl
• pH to 7.2-7.4, final volume to 1L

2. What is the differential pressure?

The core stream is where the action occurs. The cells, contained in the core stream, are spread out along the flow axis until the cells become single file as they pass by the laser intercept.

Since the sheath flow rate sets the speed of the system, the only way to increase the number of events seen by the flow cytometer is to increase the differential pressure between the sample and the sheath fluid.

The consequences of increasing the differential pressure include:

• Increasing the number of coincident events
• Increasing the spread of the data

In the data below, cells were run at three differential pressures, from low to high. As you can see, increasing the differential pressure increases the number of events (intensity increasing from left to right), but the spread of the data also increases.

Best practice is to consider running a low differential pressure at a higher concentration.

3. Does the data show backpressure or a clog?

On multi-laser systems, knowing the order of the lasers is a good thing. One impact of things that affect the sheath flow rate (i.e. clogs and back-pressure) is that the time to travel between lasers is impacted. Thus, as the signals are matched by the delay electronics, the resulting data will be wonky.

As shown here, this was the result of a problem with the sheath flow on a 4-laser instrument, and the green laser was the 4th laser in order.

Notice how the signal on the ‘poor flow’ plot bounces around. This indicates a major issue with fluidics and if this is observed, it is time to stop and do a quick cleaning/check of the system.

Remember, filtering samples is always a good idea!

4. Is the cytometer fluid pathway clean?

Cleaning the flow cytometry fluidics pathway is a thankless task. Each vendor has their own recommendations as to how often to clean the system, and what reagents are best to use.

A long clean of the system should occur at least once a week (more if it is a heavily used system), and a shorter cleaning should occur daily before use. This is in addition to any cleaning that is done between users.

In long cleaning, one should bypass any in-line filters, so that the cleaning solutions do not compromise the filter status. This process takes about 1.5 hours and uses a detergent (Contrad), an alcohol (Ethanol), and sheath fluid.

• 1% Contrad 70 for 15 minutes (in sheath tank and Sample Injection Port)
• 70% Ethanol for 15 minutes (in sheath tank and Sample Injection Port)
• Water for 30 minutes (in sheath tank and Sample Injection Port)
• Sheath fluid for 10 minutes (in sheath tank and Sample Injection Port)
• Run QC particles

One important thing to remember is that when bleach is used, it is critical to wash out the bleach before opening the machine for general use.

In this experiment, 10% bleach was run for 5 minutes on the SIP before removing the tube and placing a tube with water on the SIP.

The system was either run for 5 minutes (blue line) or not run at all (red line). Peak 6 beads were run and 10,000 single events recorded. The data shows that in the presence of residual bleach, the APC signal decreases by 50%, while the PE signal is relatively robust (only a 2.5% decrease).

A quick fix for this is to put a fresh tube of water on the SIP and start running the system while you set up the electronics — and this issue will be avoided.

As a side note, there are some other fluidics arrangements out there — such as the Attune, which uses an acoustic wave to focus the cells in the center of the core stream, and the Guava instruments, which use a microfluidics capillary system, meaning no separate sheath fluid.

These types of issues can also arise on these systems, so watch your data.

Understanding the fluidics system and observing the consequences of the system during acquisition is important to solve problems before they become major issues. While walk-away options exist on cytometers (such as high throughput sampling systems), care must be taken to make sure that the samples are properly prepared so that clogs and other preventable nuisances are avoided and data is not lost.

To learn more about how to troubleshoot the flow cytometer fluidics system and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

The power of flow cytometry is in the ability to analyze thousands of events rapidly so that finding the proverbial needle in the haystack of heterogenous cells can be done relatively easily.

The discoveries that flow cytometry has enabled because of this ability cannot be underestimated. Prior to flow cytometry, finding rare cells, let alone studying them, was a massive undertaking.

Consider the comparison (and warning) that the great inventor Nikola Tesla once expressed…

“If he had a needle to find in a haystack, he would not stop to reason where it was most likely to be found, but would proceed at once with the feverish diligence of a bee, to examine straw after straw until he found the object of his search… just a little theory and calculation would have saved him ninety percent of his labor.”

Flow cytometry provides one with the theory, calculation, and procedure they need to find a needle (rare cell population) in a haystack (whole blood, cell sample, etc.).

If you’re getting into rare event analysis, there are some additional considerations when attempting to analyze a population that is 0.1% (or lower) of the starting population that don’t apply when trying to purify a population of 10% or so.

Here are 4 keys to sorting rare cells by flow cytometry…

1. Reverse engineer your experiment.

When starting to plan for a rare event sort, it is important to consider the downstream application and begin the calculations to see how many cells you need to start with.

For example, if the downstream assay will need 10,000 cells to start with, and the population of interest is 0.02%, one will need to start with 50 million cells initially.

If one assumes a 50% recovery, the starting population becomes 100 million cells.

Continuing on with the math, sorting at 10,000 events per second, this sort would take about 10,000 seconds, or about 2 hours and 47 minutes — if nothing goes wrong, and not counting the setup time.

Table 1: Time (in minutes) to sort the number of cells, given a sort rate of 10,000 events/second.

2. Enrich your flow cytometer sample when possible.

Time on the cell sorters is often expensive.

Additionally, the longer the time it takes to get the sorted populations back, the longer one has to stay to complete any downstream processing.

This is where an enrichment step becomes quite valuable.

Consider using depletion as a method with a magnetic bead.

This process requires identifying a target(s) that is present on non-target cells.

Cells are labeled with antibodies against that target conjugated to a magnetic bead.

The whole solution is exposed to a magnetic field, and labeled cells stick to the side of the tube (or column, depending on the protocol).

The cells of interest are recovered in the supernatant.

Figure 1: Illustrating the workflow for magnetic enrichment.

Looking at the math, with our 100 million cells, a depletion of 90% of the non-target cells will yield 10 million cells.

This also enriches the target population from 0.02% to ~0.2%, while reducing the time on the sorter from ~3 hours to something in the order of ~15 minutes.

Most magnetic enrichment processes take an hour or less.

So the whole process saves over an hour!

3. Avoid common antibody panel design problems.

The more complex the panel, the greater potential for problems in the downstream sorting process.

This is especially true for rare event analysis.

The first level of panel design focuses on pairing the bright fluorochromes with low or unknown expression targets.

If this is overlaid with the rare events, brighter is better.

The other consideration in panel design is to minimize the spread of error due to spectral spread.

To do this, one must optimize the instrument to determine the error contribution in each detector based on the fluorochromes being used.

The data below comes from optimizing a BD LSR-II using the protocols outlined in this paper from Nguyen et al. (2013), Cytometry A 83:306.

To simplify, two columns of data are presented, the left showing the amount of error a given detector receives, and the right showing the amount of error a given fluorochrome contributes to the panel.

Figure 2: Measurements of the error contributed by different fluorochromes and error received by different detectors on an 18-fluorescent channel BD LSR-II

This data shows that the Blue-A detector (PerCP-Cy5.5) receives the most error of all detectors, so not an ideal choice for making a sensitive measurement.

The Blue-B detector (FITC), on the other hand, receives very little error, making it a good channel for sensitive measurements.

On the right side, PE-Cy5 is not a good choice either, as it contributes a lot of error to the panel, while FITC again contributes a much smaller amount.

Of course, drilling into the data will reveal additional information as to which detectors are more affected by a given fluorochrome.

Taken in total, this data is critical for designing a panel, especially so for rare event analysis.

Other panel design considerations have to include the addition of a viability dye.

Cells die in the process, so removing them from the sort is important.

Viability dyes have been discussed extensively in this blog, so the reader is referred to these articles on improving cell viability, reagents for identifying dead cells, and tips on sorting.

Likewise, optimizing the panel reagents by titration is a critical step to reduce nonspecific binding, and thus sensitivity.

One can also optimize the voltages using a similar method.

In this case, cells are stained with the optimal amount of antibody and a voltage range, starting below the published values of the PMT, is performed.

Data, as shown below, is plotted and the optimal voltage is identified.

Figure 3: Results of voltage optimization for two different fluorochromes.

4. Follow the four rules of flow cytometer optimization.

In addition to the information above, details about the electrics of the sorter are an important consideration.

Rule 1: The sort nozzle should be 4-5 times the diameter of the largest average cell diameter that will pass through the instrument.

Rule 2: The size of the nozzle will dictate the pressure of the sort (larger nozzle = lower pressure), which in turn will affect the frequency of droplet generation (larger nozzle = lower frequency), which in turn affects the number of cells that can be sorted (larger nozzle = slower sort rate).

Figure 4: Relationship between nozzle diameter, sheath pressure, and frequency. From Arnold and Lannigan (2010) Current Protocols in Cytometry 1.24

Rule 3: The event rate should be ¼ of the frequency.

Since what is being sorted are droplets, it is important that the correct droplet be sorted.

To do that, it is often ideal to sort a two-drop window, which means that we don’t want a second (contaminating) cell in the second droplet.

If one looks at the Poisson distribution for the number of cells in a drop for various ratios of cells to drops, one finds that data that looks like this (courtesy of Dr. Rui Gardner, head of the Flow Cytometry Core Facility at Memorial Sloan Kettering Cancer Center):

Figure 5: Relationship between number of events in a given drop with different cell/drop ratios.

Increasing the frequency to 1 cell per 2 drops will increase the speed of the sort, but also increase the chance of coincident events (2 cells in one drop), as well as the chance of two cells in successive drops.

Decreasing to 1 cell per 6 drops slows the sort down without an appreciable improvement over 1 cell per 4 drops.

Speed also increases the spread of the data, thus decreasing the resolution of the signal.

Slower is better.

Rule 4: Pulse size matters.

The electronic pulse is a sum of the cell size and beam height.

If we were to take a theoretical 5 micron cell and a cell sorter with a 20 micron beam height, the pulse width would be 25 microns.

The velocity for this sorter is 30 meters per second, resulting in a pulse time of 0.83 microseconds (calculations courtesy of Ryan Duggan).

During this time, if a second cell enters the interrogation point, it must be discarded (‘aborted’), as the electronics cannot process the signal.

The faster the event rate, the greater the chance of these aborted events.

In addition to this processing time, the system can be programmed to interrogate a longer window.

This has the advantage of gathering more light with the consequence of possibly increasing these aborted events.

On the analyzer, it may be less critical for aborts, but in the case of a rare event sort, we want to minimize these aborts, therefore it is critical to watch the abort rate on the sort counter and reduce event rate to decrease this abort rate.

Before charging ahead, you may want to use this chart, prepared by Dr. Rui Gardner, to discuss and plan any sorting experiments.

It covers many more topics that will come up in preparing for such a sort.

Figure 6: Topics to cover when considering cell sorting experiments. Prepared by Dr. Rui Gardner.

In conclusion, journeys into the realm of rare event analysis and sorting are fraught with peril. Poorly designed panels, failure to plan for the end-results, and not taking into account instrument characteristics, all can result in a failed sort, lost opportunity, and delay in the necessary data. Take the time at the beginning, before starting down the path of rare event sorting, to understand the different issues that will potentially impact your outcome and develop a plan to address each of them. With that forethought, it will become possible to identify the needle in your cellular haystack consistently and reproducibly.

To learn more about how to use flow cytometry to analyze rare cells within heterogeneous samples and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Jennifer Snyder-Cappione, Ph.D.

Have you ever wondered what has gotten into your T cells?

Are they activated, resting, tolerant, senescent, anergic, exhausted, or maybe just having an off day?

During short-lived immune responses, such as responses to acute infections or vaccines, there are three classically defined T cell states:

1. naïve,
2. activated memory, and
3. resting memory.

Here’s how to determine whether your T cells are naive, activated or resting memory, or exhausted, and how effector function plays into that.

1. Naïve cells.

Naïve T cells have not encountered an antigen and secrete IL-2 and some chemokines. However, antigen-experienced (memory) T cells, have differentiated into effector cells that can produce cytokines besides IL-2, such as IFN-g, IL-4, and IL-17.

Markers used to identify naïve T cells include CD45RA and CD62L in human and mouse samples, respectively, with CD45RO (human) and CD44 (mouse) present on memory T cell populations.

It is best to use functional profiling as the primary determinant of T cell state in these instances, with surface markers as supporting evidence. Therefore, defining naïve T cells by their inability to exert effector functions such as IFN-g, IL-4, and IL-17 in response to either specific antigens or other stimulations (e.g. mitogens, anti-CD3/28 antibodies) is recommended.

2. Memory cells.

T cells that express CD45RO/CD44 (human/mouse) and do exert effector cytokines can be defined as memory cells. To distinguish between an activated and resting memory state, a key difference is their polyfunctionality ex vivo, with true resting memory cells typically able to secrete a panoply of cytokines, and activated cells not necessarily able to do the same.

Also, activated cells will likely express some activation markers (CD69, CD25, etc.) and show evidence of active proliferation (Ki67+, for example) that may be absent on a true resting memory T cell.

During chronic diseases, such as HIV infection, cancer, and autoimmunity, it becomes increasingly difficult to define the state or “mood” of the T cell.

T cells, (like human beings) are significantly influenced by their environment and, with the complexities of cytokine milieus and other dynamic influences in chronic inflammatory states, figuring out the exact state of the T cell can be challenging.

3. Exhausted cells and the “others”.

There is much ado of late in immunology about “exhausted” or “checkpoint inhibited” T cells.

Exhausted T cells are believed to have progressively lost their effector function capabilities due to prolonged antigenic exposure.

Exhausted cells have a distinct molecular signature from naïve and traditional effector cells, and they are often defined via hypo-responsiveness to stimulation ex vivo and expression of inhibitory receptors (IRs), such as PD-1, TIM-3, and CTLA-4, with individual T cells expressing multiple IRs possibly suggesting of a more exhausted cell.

However, activated T cells can also express IRs, so it’s best to define IR+ hypo-responsive cells as ‘bearing a phenotype resembling exhaustion’ and if one wants to define the exhaustion phenotype more definitively, comprehensive molecular signatures should likely be measured.

Other terms to define memory T cell states of lower functional capacity include anergy, tolerance, and senescence.

Anergic cells are classically defined as non-responsive due to insufficient priming in vitro. Similarly, tolerant T cells are also improperly primed as they are self-reactive.

Senescent T cells are associated with aging and characterized via shortened telomeres and expression of CD57.

To date, the lines distinguishing senescence from exhaustion are blurred.

In summary, it’s suggested that you define your T cell population of interest by its functional profile (1^st), surface marker expression (2^nd), and finally take into account what you know about the cells (for how long have they been in an inflammatory, antigen-rich environment?). From this combined information, speculate as best you can on the T cell population’s status. Until T cell differentiation and effector function regulation is better understood, we can only make reasonable guesses as to the state of many T cell populations.

To learn more about how to differentiate T Cell state with flow cytometry and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

You never forget your first experiment.

Usually, you remember all the things that went wrong that you need to correct for future experiments.

In flow cytometry, there are so many things that you need to remember to bring to the first experiment, that having a checklist make sense. This checklist is divided into three phases — before the experiment, at the instrument, and after acquisition — with steps listed for each to help you perform your first flow cytometry experiment with ease.

1. Preparation before the Flow Cytometry experiment.

Before embarking on the first flow experiment, there are several things that you should do to become comfortable with the experimental plan. The first is to understand the protocol that will be used to stain the cells.

1. Meet with the flow cytometry staff. One of the first steps to take before planning your first experiment is to meet with the team that manages the flow cytometer you will be using. These people are a great reference on the ins and outs of flow cytometry experimental design, execution, and analysis. You will probably have had training on how to acquire data, but a one-on-one meeting to discuss your assay and get feedback is a great way to learn more before embarking on the experiment.
2. Sample preparation. For those starting with a liquid sample (pond water, blood, suspension culture cells), it’s pretty easy to get to the single cell suspension. When working with solid tissues, dissociation techniques become critical and references like the Worthington Tissue Dissociation Guide are excellent. Of course, don’t forget to filter samples before running them through the flow cytometer, especially if you are planning a sorting experiment. Finally, make sure to take a look at your cells under a microscope — it is your best friend to assess the quality of your single cell suspension.
3. Gather and validate reagents. Make sure you have everything you will need, and ideally the reagents have been tested already. It really is frustrating to start a new experiment with a new technique using a reagent that you don’t know will work. New members in a lab might start with staining a standard sample with standard reagents until you are comfortable with all steps of the process. This sample is the same one used as a reference control in the assay, meaning there is a great deal of experience on how it should perform. Don’t forget to know any special issues with reagents — for example, don’t leave tandem dyes out in the light or on the bench for any length of time, lest they degrade and become useless.

Figure from Derek Davies.

4. Book time. Make sure that you add some extra time as you get used to using the software. If you met with the staff already, ask them for an idea on how long they think the acquisition will take. Of course, don’t forget to add time to perform the necessary cleaning at the end of your acquisition.

2. At the cytometer.

Now comes the big day — the first experiment.

As with everything you do for the first time, it will take longer than you expect. Patience and care to follow the protocol are essential for success. Take a deep breath and begin.

1. Double-check the necessary controls. Flow cytometry experiments require a large number of controls for successful interpretation. These include:
   1. Compensation controls — for setting compensation
   2. Fluorescence Minus One (FMO) controls — for assisting in gating
   3. Biological controls — for addressing variation in the assay
   4. QC controls — usually beads that will be used to monitor the QC of the assay over time

   Forgetting a control, especially in the first experiments, makes the downstream analysis harder. These controls each play an important part in the analysis, and forgetting one can mean success or failure. So double-check the list.

2. Count your cells. Making sure you have enough cells for your experiment is important. Nothing ruins the day like realizing you don’t have enough cells to do what is needed. There are a host of ways to count cells,
3. Make sure you have your cheat sheets. Let’s face it, the checklist you got when you learned how to run the analyzer is important. It will help guide you through the process of using the software (a new skill) and ensure you don’t forget an important step (like hitting the “Record” button to save your data). Refer to it early and often. Soon, it will be like riding a bike, but for now the cheat sheet is your training wheels to get going.
4. Arrive on time, if not a bit early. First-time jitters are common, so getting to the instrument a bit early doesn’t hurt. Being late, however, does. If you are early, let the staff know and see if anyone has time to be ‘on-call’ should you have problems. Also, make sure you transport your samples from your lab to the flow cytometer in an appropriate manner. Coolers like these from Igloo are a good choice because they are top-loading, have a closing lid (that ‘locks’) and provide a secondary containment for your samples. They also protect the sample from light.
5. Annotate your data. Get in the habit early of using and entering keywords for all your experiments. In many acquisition packages, you can add keywords at the experiment level, the sample level, and the tube level. Using keywords and making a habit of it is something that will be its own reward when you are analyzing your data hours before a deadline. There are many things you can do with properly annotated data in third party software as well, so make sure you take advantage of this from the beginning.
6. Follow the protocols. Check for the recommendations for best voltage, and how to collect your data and export it for later analysis. Make sure you collect enough events. Some packages default to a very low number of events, so check to be sure you are collecting enough data. At this point it’s time to run the samples and collect the data until the last tube is completed.

3. After acquisition.

There — the easy part is done. The first experiment is completed and in the books. Your data is being exported and saved to the place you’ve been told to save the data. Now what?

1. Clean the machine. Follow the protocols for cleaning the instrument before the next user. Being a good citizen helps make core facilities run smoother, causes fewer issues for the next user, and in general is a good thing to do. Did you know, for example, leaving bleach on the sample injection port (SIP) can kill the next user’s fluorescence?

So follow the cleaning protocols — your fellow users will thank you. The core staff will thank you.

2. Double-check the data is exported. Make sure the data is properly stored and you will be able to access and analyze it later. Sometimes there are hiccups in exporting the data, so checking to make sure each sample was exported (and there is data there) will save you a trip back to the instrument to get the data off a second time.
3. Breath deep and prepare for analysis. Staining and running the experiment on the flow cytometer is the easy part. The hard part — and in many cases the fun part — is about to begin. That is the analysis of the data. Here, you will encounter your own set of new problems: from learning the analysis software, fighting the urge to ‘tweak’ your compensation matrix, making sense of the data, putting the gates in the right place, and so on. We’ll save your first analysis for another time. For now, breath deep, make sure everything is cleaned up, and enjoy that feeling from a first experiment.

You did it. You have survived the first of many flow cytometry experiments. Flow cytometry is a very powerful tool and can answer many questions if the experiments are properly designed. Using this checklist will help you design and perform consistent experiments every time. There is a learning curve that takes a bit of time, patience, and practice, but soon you may be finding excuses to perform flow cytometry experiments and we will be here to help you with best practices.

To learn more about how to prepare for a flow cytometry experiment, and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

“If one approaches a problem with order and method there should be no difficulty in solving it — none whatever.”

— Hercule Poirot, Death in the Clouds

Murder is a common theme in the mystery suspense genre. The detectives who solve these murders use a combination of observation and deduction to identify the guilty party. This metaphor suits measuring cell death.

In biology, there are four major pathways for cell death.

The study of the different ways cells die has become known as Cell Necrobiology, as coined by Darzynkiewicz and coworkers in their 1997 review article.

Flow cytometry is ideally suited as a tool to study Cell Necrobiology and, with its plethora of reagents, it is even possible to follow the different steps in these processes.

The four major ways a cell can die are:

• Apoptosis — programmed cell death
• Necrosis — traumatic cell death
• Autophagy — ordered degradation of cellular components (autophagocytosis)
• Necrotosis — ordered necrosis

Cell death is so important, that it has been the center of several Nobel prizes including one awarded in 2002 to Sydney Brenner, Robert Horvitz, and John Sulston who discovered the genes involved in apoptosis, and again in 2016 when Yoshinori Ohsumi was recognized for his work on the mechanisms of autophagy.

Of these mechanisms, apoptosis is probably the most readily studied using flow cytometry.

There are many assays that can be performed to measure apoptosis in cells. These can be grouped by the state of the cell as it dies.

1. Measuring apoptosis.

One of the first stages of apoptosis are changes seen in the mitochondria, where the membrane potential collapses, which leads to the release of several factors that can inhibit anti-apoptotic proteins, as well as the release of cytochrome c, which binds to another protein that ultimately causes the activation of caspase-9, and in turn caspase-3.

There are a host of dyes that can measure this depolarization, including CMXRos, JC-1, and TMRE.

One of the most common is JC-1, which is a cell permeant dye that emits a red fluorescence (~590 nm) in healthy, active mitochondria because of the formation of aggregates. As the membrane potential collapses, the aggregates fall apart and the fluorescence shifts to a green color (~529 nm).

This is observed by flow cytometry as a change in the ratio, as shown in this data from Derek Davies, head of the flow cytometry facility at the Francis Crick Institute in London.

Here, untreated cells show a dominant red fluorescence, but after drug treatment, there is a dramatic shift to green fluorescence.

The next steps in apoptosis include the activation of the caspases, and changes in membrane symmetry and permeability.

Phosphatidylserine (PS) is found on the inner leaf of the plasma membrane. As apoptosis progresses, PS flips to the outer membrane, which is a signal for the cells to be phagocytosed.

The protein Annexin V is a calcium-dependant protein that preferentially binds to PS. When you add a cell-impermeant dye, such as 7AAD or PI, you get a very robust assay for looking at apoptotic and necrotic cells. Typical data are shown below:

Annexin V is calcium-dependant and not very stable. In general, it is best to read Annexin-stained cells within an hour or so of staining.

If you’re planning to perform a lot of Annexin assays, this buffer works very well:

10x Annexin Buffer

0.1 M HEPES

1.4 M NaCl

25 mM CaCl₂

You can stain cells with surface markers, which is best done before staining for Annexin V.

2. Measuring necrosis.

One of the hallmarks of necrosis is the loss of membrane integrity, which leads to the easy use of a host of cell-impermeant dyes from PI to 7AAD and others.

In the Annexin assay above, cells that are Annexin negative and DNA dye positive are often considered to have died by necrosis. Unfortunately, these cells can also show up in the Annexin positive, DNA dye positive fraction, making it an imperfect measure of necrosis.

It turns out that high-mobility group B1 protein (HMGB1) may be able to differentiate necrotic cells.

This nuclear protein stays contained within the nucleus during apoptosis, but is released when cells undergo necrosis (Raucci et al., 2007). Currently, this is typically done by analyzing the supernatant for HMGB1, or by microscopy.

This would be an excellent assay to implement on a tool like the ImageStream.

3. Measuring autophagy.

Historically, autophagy has been defined by the measure of the ‘autophagosome’ by either electron or light microscopy.

The marker LC3 could also be used in microscopy techniques, especially since LC3 was best measured when it was tagged with a fluorescent protein like GFP.

Fortunately, a paper was recently published by Chikte and co-workers (2014), in which the authors report the use of the dye Lysotracker Green DND-26 as a way to measure autophagy in cells.

They compared this dye to LC3 and demonstrated that Lysotracker gave similar results, and importantly, was easier to use than LC3.

Using flow cytometry and a host of different reagents, it is possible to tease out how your cells may have died. Like the most famous consulting detective once said, “When you eliminate everything else, that which remains, however improbable, must be true” (Sherlock Holmes, the Sign of Four). Using these tools, you can readily eliminate the various suspects and come to your conclusion as to how your treatment may have killed your cells of interest.

If you’re interested in further reading, this review article by Wlodkowic and coworkers (2011) Methods Cell Biol 103:55-98 is an excellent reference.

To learn more about how to use Flow Cytometry to measure Apoptosis, Necrosis, and Autophagy, and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

How exactly does a photon of light become an electron, and eventually a number in the FCS listmode file?

In flow cytometry, cells are labeled with fluorescent tags, passed by an excitation light source, and the resulting emitted photons are collected to become the data that identifies characteristics about the cells in question.

These characteristics can include what proteins are expressed, the cell cycle state, the phosphorylation state of a protein, the calcium state of the cells, levels of RNA expression, and so much more.

If there is a fluorescent probe that can measure a specific characteristic, it can often be used in flow cytometry.

A great resource for reference is the Molecular Probes Handbook, as it contains a wealth of information about different fluorescent reagents that can measure everything from apoptosis to Zn++ ion concentration.

This process involves the detection system and the electronics, and basically following a bouncing photon to its ultimate digitization.

The role of the detector on a flow cytometer is to capture photons and convert them to electric current.

By and large, the two most common detectors on commercial flow cytometers are photodiodes and photomultiplier tubes.

In a PMT, photons of light enter through a window and hit the photocathode. If the photon is of sufficient energy, they eject an electron (due to the Photoelectric Effect).

The electron is focused to one of several ‘fins’ (called dynodes), where the electrons are multiplied via secondary emission. At the end of this chain is an anode, where the electrons are converted to an electronic pulse. This is shown schematically below.

As is shown, there is a linear relationship with the photons hitting the cathode and the output photocurrent.

It should be noted that PMTs can turn any photon that strikes the photocathode into a photocurrent.

Thus, to control the photons that are measured by any given detector, optical filters are placed in front of the PMT.

Most commonly, these are band pass filters, and as a reminder and shown below, band pass filters allow light within a given range through. To describe these filters, manufacturers label these filters with the center of the range and the size of the window of light that can pass through the filter. So a 620/40 band pass filter will allow light between 600 to 640 nm through, as shown below.

Current produced is proportional to intensity of the fluorescent signal:

At the end of the day, one has a current produced that is proportional to the intensity of the fluorescent signal, and each PMT measures a specific wavelength range because of the optical filters placed in front of the detector.

At this point, the photons of light have been turned into electric current.

The next question is, “How does that electric current get turned into a digital value that is sorted in the FCS file?” The image below shows what a typical electronic pulse looks like.

This pulse has three characteristics that can be measured: the peak height, the pulse width and the pulse area, and the integral of the height and width. The boxes represent the fact that the electronic pulse is being sampled at some frequency, and with each sampling, the values are digitized by the Analog to Digital Converter (ADC).

The ADC has two characteristics: the sampling frequency and the resolution.

The sampling frequency is expressed in megaHertz (MHZ), and ranges from 10 to 100 on the current generation of instruments. This means that the electronic pulse is being sampled between 10 to 100 million times PER SECOND.

From this sampling, the value is digitized into a range equal to 2 to the power of the ADC. This can range from 10 to 24 bits or 1,024 to 16,777,216 discrete values.

The values for some common instruments are shown below:

It is worth noting that BD instruments use a predetermined 18 bit log-lookup table so that the range of the data is 262,144 bins.

Frequency of sampling and resolution in a Flow Cytometer:

There are two important characteristics of the ADC: the frequency of sampling and the resolution.

Natural questions might include: What are all these bits, really? Or are they bins? Or channels?

In the case of a 10 bit ADC, this means there are 1,024 discrete values that the sampled pulse can be assigned. With a higher ADC (Accuri with a 24 bit ADC), it has over 16 million discrete values.

These are the ‘channels’ or ‘bins’ and represent the actual value that has been measured from the signal pulse. So when the signal is digitized, it is assigned one of these values.

For linear data, this is pretty self-explanatory. However, with log data, this binning gets a bit tricky, especially if you consider the difference between the older FCS2 data and the current FCS3 data.

Take, for example, the FACSCalibur, which has 10 bit resolution. Since this data, when log transformed, was being measured over 4 decades, the vendor decided to spread the 1024 bins equally across the log space.

So the first decade (1-10) had 256 values, as did the second decade (10-100), the third decade (100-1000) and the fourth decade (1000-10,000). One can see that in the case of the third and fourth decades, 256 bins did not give sufficient resolution, so that each bin contained more ‘values’ than a bin in the lower decades.

On the modern digital instruments, these values are spread out to better reflect the number of values available at the higher level ADCs. Therefore, there is more resolution at higher decades than at lower decades.

Flow cytometrists tend to use bins and channels interchangeably.

Conversion of photons to electrons:

The number of bins that a given system has is based on the resolution of the ADC.

With digital instruments and the FCS 3 format, the bins are spread out to better reflect the number of possible values in the log space.

In summary, a photon is emitted from a fluorochrome. The photon is captured by the detector (e.g. the photomultiplier tube, PMT), where it is converted from light to an electron. At the same time, the PMT amplifies the incoming signal in a linear, proportional manner.

The output is an electronic signal pulse. The electronic pulses coming off the PMT is sampled many millions of times a second, and the height of that sample is digitized into a discrete value that is placed in a bin. The resolution of the ADC dictates the number of possible bins available for use.

In the end, this value is stored in the FCS file, and ready for further analysis.

To learn more about how flow cytometry converts photons to digital data, and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

Controls are critical for minimizing the effects of the variables in a scientific experiment so that the effect of the independent variable can be accurately measured.

In flow cytometry, there are a host of important controls necessary to properly interpret data generated in these experiments. Some of these controls include compensation, fluorescence minus one (FMO), stimulated and unstimulated, reference, and controls.

When it becomes time to publish, the proper use of these controls is critical in convincing the reviewer and reader that the data has been properly analyzed.

The isotype control is an experimental control where a sample is stained with an irrelevant antibody with the same isotype as the target antibody. Cells are gated and positivity is set based on the background staining of this isotype control.

The use of the isotype control to set negativity remains a topic of discussion and can confuse the novice to flow cytometry, especially when a reviewer may request why these controls were not included in a submitted paper.
Overall, the isotype control is one that is often overinterpreted and can provide little additional information in an analysis.

As Keeney and co-workers stated in their 1998 paper, “Isotype controls in the analysis of lymphocytes and CD34+ stem and progenitor cells by flow cytometry — time to let go!”

(Author emphasis).

Clearly, the use of the isotype control is ingrained in the scientific community, and vendors are more than happy to sell you one (or more).

Delving deeper into this specific control and evaluating the strengths and weaknesses can help you to come to your own conclusion on if and when to use an isotype control in your flow cytometry experiments.

1. What is an isotype?

As a B cell matures, it undergoes V(D)J recombination, which results in the production of the B cell receptor (BCR). After selection in the bone marrow, the B cells circulate to secondary lymphoid organs, where they are constantly exposed to new antigens.

When the BCR binds an antigen, the B cell becomes activated and can secrete antibodies as well as generate a memory cell, to provide long-term protection.

Class switch recombination occurs in mature cells upon stimulation, and in the presence of signaling molecules. The variable region of the antibody is retained, while the constant region of the heavy chain is changed, based on the specific signal molecules present at the time. This process is illustrated below.

From: Wikimedia.org

Thus, the isotype is changed.

2. What is an isotype control?

An isotype control is an antibody to an irrelevant target that shares the same heavy and light chain as the target antibody.

For example, the anti-human CD3 antibody HIT3a has the isotype of Mouse IgG2a, κ. Thus, one would look for an isotype control with the same characteristics.

There are several assumptions that are made when an isotype control is used:

1. There is no target for the isotype control antibody expressed on the cells of interest.

   The problem with this assumption is that the target of the isotype control is not always known. For example, one can purchase Mouse IgG2a, κ, clone MOPC-173 as an isotype control for CD3 (clone HIT3a). However, the antigen details of the MOPC-173 clone, as described on the BioLegend website (and many other vendor’s sites) states:

   Thus, the target is not known, it was only selected as an isotype control because of screening against various samples.

   This does not mean that the target is not expressed on your specific cells of interest, but only that it doesn’t bind in common tissues.

2. The non-specific binding of the isotype control has similar characteristics to the target antibody.
   There are three factors to consider in antibody binding:
   • Specific binding — this is the binding of the antibody to the target of interest. This is what we are interested in.
   • Fc receptor mediated binding — this is a specific binding of the constant region of an antibody to the Fc receptor expressed on certain cell types. Generally, this is not something we are interested in, and is usually dealt by various blocking methods.
   • Non-specific binding — this is off-target binding of the antigen to any protein in the cell. If the antibody can’t find the target, there is a chance it will bind to another protein. This is often driven by antibody concentration, and one of the critical reasons for titration.

   Since we don’t necessarily know the target of the isotype control, it is impossible to know what the off-target binding will be.

   We are left with the fact that the isotype antibody has been ‘tested’ against a standard series of cells and cannot be sure that the NSB of the two reagents will be similar.

3. The fluorochrome to protein (F/P) ratio is the same between the isotype control and the target antibody.

   The F/P ratio represents the amount of fluorochrome that is bound to the antibody. In some cases (for large fluorochromes like PE and APC), this is typically in the 1/1 range. However, for smaller fluorochromes, this is not the case, and the labeling of each antibody must be optimized. Thus, unless you know the F/P ratio of both the target antibody and the isotype control, differences in staining and fluorescence could be due to an F/P difference.

3. Should I use isotype controls?

That is the real question.

In addition to the Keeney paper mentioned above, there are several papers that suggest to not rely on isotype controls, including Maecker and Trotter (2006) and Hulspas et al., (2009).

Most recently, an excellent article by Andersen and coworkers (2016) explored the question of how best to block nonspecific binding. Their work was performed on monocytes and macrophages, which contain large amounts of Fc-Receptors.

Figure 1 from Andersen et al., (2017)

This is Figure 1 from Andersen’s paper showing how, in the absence of blocking, the isotype control (IgG1) binds at a much higher level than the specific binding for Tie2, a protein known to be expressed at low levels on monocytes.

However, the plot gets more interesting as the authors explore how best to block this Fc mediated binding of Isotype controls, which they further demonstrate was specific to the Monocyte subset, which is shown in the third figure of this paper. Interestingly enough, the authors demonstrated that this nonspecific binding was only seen with the IgG1 and IgG2a isotype controls.

Figure 3 from Anderson et al. (2016).

The final piece from this paper is the report that different lots of the same isotype from the same vendor showed different binding responses. Again, calling into significant question the use of isotype controls in the setting of gates to determine positivity of a given antigen.

The conclusion from this paper nicely sums up the best practices researchers should be using — and it doesn’t include isotype controls. To quote from Andersen’s paper:

There is one, albeit small, use for isotype controls, also illustrated in the Andersen paper. An isotype control can be used to show that there was poor or incomplete blocking of specific subsets being labeled. That is it.

Looking at all the data and discussions, from the 1998 Keeney paper to the most recent work by Andersen, it is clear that the hypothesis that an isotype control can show the background or nonspecific binding on cells must be rejected. If a reviewer rejects your paper because of the lack of isotype controls, you now have the needed information to rebut those arguments. Likewise, if someone is using an isotype control to set background staining levels, this information can help them realize that they are wasting time and money by using this control for that purpose.

To learn more about the strengths and weaknesses of isotype controls in flow cytometry, and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

When a researcher first sits down in front of the flow cytometer, they are faced with several choices.

One of the most confusing is, “What voltage do I set my detector at?”

Unless the researcher is running a fixed voltage system (Accuri, and others), this choice can dramatically impact the sensitivity of the instrument, making or breaking the experiment.

In the days of the analog flow cytometers, voltages were set by placing a quadrant gate on a bivariant plot, with the lower left quadrant encompassing the first log decade. Unstained cells would be run on the instrument and the PMT voltage set until these cells were contained within that lower left quadrant.

Figure 1: Schematic of setting voltages ‘Old School’.

Because of how the data was processed in older generation instruments, a portion of the population was ‘off-scale’ and accumulated in the first channel.

Using this method, the highly autofluorescent cells would drive the voltage, often causing a compression of the less autofluorescent cells on the axis. However, this was not always obvious because of how the data was plotted.

It should be remembered that with these systems, the data was log transformed (and compensated) in the hardware, and stored that way. This prevented much post-acquisition manipulation.

Enter the digital cytometer age, in which the data is processed post-acquisition.

This paradigm change in data acquisition and storage also meant that traditional methods for setting voltages needed to be reviewed and new methods developed.

1. Peak 2 optimization.

All PMTs have a sensitive sweet spot: a voltage at which the detector is most sensitive for the fluorochrome emission being measured.

The first paper to discuss a practical application of determining the voltage sweet spot was published by Maecker and Trotter in 2006. In this paper, the authors discuss what is termed the ‘Peak 2’ method for determining sensitivity.

In this method, a very dim particle (peak 2 of the Spherotech 8 peak bead set) is run over a series of voltages, and the Coefficient of Variance (CV) of the bead is plotted versus the voltage.

A typical graph is shown in Figure 2.

Figure 2: Results of a Peak 2 optimization of a PMT.

As PMT voltage increases, CV of the beads decreases until it hits an inflection point, and there is no improvement in the CVs from that point. The red arrow indicates the ‘optimal’ sensitivity.

This graph is interpreted by identifying the inflection point on the graph (shown in the red arrow).

Increasing the voltage from this point does not improve PMT sensitivity (with a caveat, discussed later), so the best voltage to start at is just below that inflection point.

2. Cytometry setup and tracking.

BD ran with this idea, and implemented a different method in their Cytometry Setup and Tracking.

When a CS&T baseline is run, the optimal voltage is determined by first determining the Standard Deviation of Electronic Noise (SDEN).

The software calculates the voltage necessary to set the baseline at 10X SDEN. The results of this method are stored in the CS&T report, and a typical curve is shown in Figure 3.

Figure 3: Results of a CS&T optimization baseline report.

If you have a digital machine, made by anyone other than BD, what can you do?

Option one is to follow the peak 2 method, which works very well.

A second option is to measure the electronic noise in your system and generate a voltage via CS&T.

One method to measure electronic noise is to run a negative particle, (better yet, unstained cells), over a voltage range. Then, generate a plot of (1/mean² vs CV²), and calculate a regression line.

The slope of the line is the Variance of Electronic Noise, and the standard deviation is the square root of the variance. This gives you a channel value for electronic noise. Multiply by ten to get the target value for each channel.

Now put your unstained cells on the instrument, adjust to the target value, and away you go.

Figure 4: Calculating SD_EN. Results of calculating the SD_EN using PBMCs run over a voltage series.

You can improve these target values by performing a voltage optimization using the target cells stained with the properly titered reagents for the specific panel.

This can help improve sensitivity because of the fact that the spectrum of fluorochrome of interest in your panel may not be accurately modeled using the beads.

As shown in the figure below, two different fluorochromes give very different results.

Figure 5: PMT optimization.

Cells were labeled with optimal concentrations of antibodies of either CY7-APC (Left) or BV650 (Right), and a voltage optimization was performed, starting below the recommended peak 2 value.

The separation index was plotted against the voltage and the curves above generated.

Particularly with the newer dyes, an improvement in the separation index identifies a better voltage for this cell/fluorochrome combination.

3. Setting consistent target voltages.

The last step in voltage setting on a digital instrument is to have some way to consistently set those target voltages over time, without having to go through a long, tedious process.

Beads again come to the rescue. In this case, a very bright bead (some favor the 6th peak of the 8 peak bead-set, but you can choose your favorite), that is fluorescent in the channels to be used is run when the optimal voltages are identified. This results in a value for each detector — the ‘target value’.

Then set up a template with the appropriate plots and target values and remember to save it for later use!

When you come back to the instrument, open up the template and run this bright bead. Adjust the voltages (as necessary) to achieve the target values in each channel and you’re good to start acquiring your samples.

The other advantage of this method is that you can become cross-platform compatible.

The best way to take out the fear and agony of setting voltages is to use some optimization methods. The peak 2 method, as described above, is a useful and robust method of identifying optimal PMT voltage ranges. Refining that to the voltage walk with the actual cells and fluorochromes of interest will further improve sensitivity, which is especially critical for rare cell populations or emergent antigens (like activation markers). Of course, don’t forget to set-up a way to monitor and maintain those voltage settings.

To learn more about how to set and monitor optimal voltages for a flow cytometry experiment, and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

On most flow cytometers, the photomultiplier tube (PMT) is the interface between the fluidics system and the electronics system. It is the PMT that converts the photons emitted from the fluorochromes into the electronic current that is digitized and ultimately converted to the value stored in the listmode file.

Any stray photon of light or random electron emission from a dynode will cause a cascade, and ultimately a photocurrent. This is often known as dark current. The figure below shows the idealized idea behind this concept.

Figure 1: Stylized signal coming off a PMT showing the dark current and actual signals of cells passing the laser intercept.

As this figure shows, if each of these peaks is counted, there would be over 50 ‘events’ seen by the flow cytometer. Most of these events would be considered junk or debris.

Imagine if each of these events was recorded in the listmode file — how large would the file be?

To reduce this background noise in the system, we can use something called a threshold.

The threshold is a value that the signal must be above before the system will call a pulse an ‘event’.

If one enforces a threshold, the resulting pulse would look like this:

Figure 2: Stylized signal with a threshold added.

This threshold now reduces the signal from over 50 events down to just two. A much more manageable file size and analysis will be able to be performed by removing this noise.

1. Thresholding reduces background noise in the flow cytometer.

Thresholding is a powerful tool for reducing the signal caused by debris and dark current present in the flow cytometer.

Thresholding is a useful tool in reducing the debris that can overwhelm your datafile.

When computer storage was more expensive, and computers less powerful, it was much more heavily used.

Judicious use of threshold is warranted. However, it is also important to remember that if the flow cytometer doesn’t see an event, it doesn’t mean that the event is not present in the sample. This is especially true when sorting cells.

Here is an example of the effects of increasing the threshold on populations. These are CS&T beads run on a FACSAria, with an increasing threshold from 5,000 to 50,000 on the forward scatter parameter. A total of 20,000 events was collected for each file. Two gates (small and sort) are indicated.

Figure 3: Effects of increasing threshold on CS&T beads.

As can be seen by this data, increasing the threshold, decreasing the amount of debris seen in these beads, and the percentage of events in each gate, changes. These values are shown below.

Table 1: Percentage of cells in the two gates from Figure 3.

2. Thresholding removes smaller events.

Thresholding increases the percentage of target events in the datafile by removing the smaller events.

If one was performing immunophenotyping analysis, for example, the increased threshold resulting in a loss of the events in the ‘small gate’ would probably not be of concern. However, if one was to sort based on these different thresholds, a very different picture emerges.

The sort logic for this experiment is shown below, as generated in DIVA.

Figure 4: Sorting Strategy for Threshold Sorting Data.

After sorting the beads at different threshold levels, a post-sort analysis was performed on the instrument. Before the beads were placed back on the system, the threshold was reset to 5,000 and a wash was performed for two minutes, and the data from that wash shows the amount of background noise in the system.

A total of 89 events were observed in two minutes, with two of these events in the ‘small gate’ and none in the ‘sort gate’.

Figure 5: Results of a two-minute water wash showing the background noise in the system after sorting beads.

The results of the pre- and post-sorts are shown in figure 6. This is data from the ‘bright cells’ sort is shown, and the beads were sorted on purity mode.

Figure 6: Results of pre- and post-sort analysis at either 10,000 or 50,000 threshold on forward scatter. Post sort analysis was performed with a threshold of 5,000.

Table 2: Data from sort gates in Figure 6

The numbers tell the tale.

At a low threshold, the purity of the post-sort population of interest is extremely high, with only a minor contamination of the events in the ‘small gate’. However, when the instrument is blinded to the small events (via a high threshold), the post-sort analysis shows that there is significant contamination and much lower purity.

Since the system could not see the small events, it was not possible for these events to be excluded, thus the small events were sorted AT RANDOM into the collection tube, because the system did not abort those droplets where a small event was in the leading or lagging droplet.

While increasing the threshold will speed up the rate of acquisition of the events of interest, the effect of increasing the threshold must be weighed against the sensitivity of the downstream application.

If the cells are to be cultured, this debris may be tolerated. If the downstream analysis is a very sensitive technique, such as RNAseq, this debris might not be tolerated. It pays to be careful with the threshold to avoid surprises — like having your highly purified cell population contaminated with a host of unexpected genes (say 𝛃-globin).

Adding a threshold when acquiring flow cytometry data is like putting on sunglasses on a sunny day. It reduces the number of events by setting a bar that a signal pulse must clear before it is counted as an event. Depending on the importance of the data, the downstream applications for the data (or sorted cells) will dictate how critical the threshold is. Threshold wisely and practice proper sample preparation to reduce the debris in the tube to ensure the best outcome.

To learn more about how to use a threshold to reduce background noise in Flow Cytometry, and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

Fluorescence compensation is not possible without proper controls, so it is critical to spend the time and effort to generate high-quality controls in the preparation of an experiment.

First, recall that compensation is a consequence of spectral overlap, which occurs because the fluorescence emission spectra of essentially all fluorophores are wider in wavelength range than the optical filters that we use to measure those fluorophores.

Because of this, fluorophore emission will often overlap with more than one filterset on the cytometer, leading to the detection of the signal from one fluorophore by multiple detectors (i.e. spillover). The more colors measured within a single experiment, the more crowded the spectrum becomes and the more severe this kind of crosstalk is.

Compensation is a mathematical process that deals with this problem by removing a percentage of the total signal from each detector.

This percentage corresponds to the amount of signal spillover signal that is contributed to this detector by all the other fluorophores being used in the experiment. The compensation calculation relies on the fundamental concept that the amount of spillover of a fluorophore (e.g. fluorophore A) into a detector (e.g. fluorophore B) is defined by the ratio of A’s signal in B’s detector to A’s signal in its own detector.

This relationship between two detectors, the spillover coefficient, defines the spillover regardless of the amount of dye, so it can be used correctly for the spectral overlap in a typical experimental setting in which amounts of dye can vary widely between and within samples.

Given all of these points, the most important aspect of setting up a relevant, proper, and functional compensation matrix is to ensure that the spillover coefficient is accurate, is wholly reliant on good controls.

What Is A High Quality Compensation Control?

What does it mean for a compensation control to be “proper” or “good”?

The Daily Dongle, contains a much-cited and useful post, Three Rules For Compensation Controls and answers this question with three guidelines:

• Each compensation control must be as bright as, or brighter, than the experimental stain.
• Autofluorescence should be the same for the positive and negative populations used for the compensation calculation in each channel.
• The fluorophore used must be the exact fluorophore (i.e. same molecular structure) that is used in the experimental sample.

Let’s look at each of these in detail.

Keep in mind that properly compensating a channel requires both a fluorescent, or stained population (for every fluorescent channel), and a nonfluorescent, or very dim, population.

These populations can either be present:

• in a single tube (e.g. a lysed whole blood stained with a marker which will be present on only a portion of the total number of cells)
• or, in separate tubes.

In the latter case, the controls would consist of (1) a single universal negative tube, used to designate the negative population in all channels and (2) a tube containing stained cells or beads for each channel.

Note that the single-stained controls may or may not contain negative cells. Regardless, when using a universal negative, the operator instructs the cytometry software to utilize the nonfluorescent population in the universal negative control and to ignore any nonfluorescent cells or beads in the single-stain controls.

3 Keys To Creating High Quality Compensation Controls

1. Account for brightness.

You may be wondering whether the compensation matrix may be irrelevant and inaccurate if compensation controls signal intensities are not exactly matched to the experiment’s samples.

The answer is a firm no: they don’t need to be matched. This is the beauty of compensation…

The spillover coefficient defines a ratio of interaction between two channels that is independent of the amount of fluorophore.

Another way of thinking about the spillover coefficient is that it defines the slope of a line that can be drawn between the nonfluorescent and fluorescent populations, connecting their medians, in an uncompensated plot of data from a single-color control.

The slope, or angle, of this line defines the amount of compensation required to correct the spillover between the two channels on the plots (see Figure 1 below).

The accuracy of the slope of the line described above is only as robust as the quality of the compensation controls used to generate it.

This accuracy can be compromised when the signal in the experimental sample is brighter than the signal from the compensation control for a given channel.

By nature, dimly fluorescent populations tend to spread out (have higher coefficients of variation). This is because fewer photons are measured from a dimly fluorescent population than they are from a brightly fluorescent population.

When a population is spread out in this way, there is uncertainty in the determination of the median, which will introduce uncertainty into the determination of the spillover coefficient (see Figure 2 below).

2. Account for autofluorescence.

Autofluorescence is an inevitable component of the total fluorescent signal in any channel.

This contribution has no effect on compensation as long as its extent is identical between the fluorescent and nonfluorescent populations used to calculate compensation.

However, the story is quite different if the autofluorescence intensity is unequal between these fluorescent and nonfluorescent populations (see Figure 3 below).

If the autofluorescence intensity of the bright and negative populations in a compensation control is augmented identically, both populations will shift on the scale of the channel to be compensated.

Nevertheless, the slope of the line that can be drawn between their medians is unaffected; therefore, the spillover coefficient and the compensation matrix will also be unaffected.

However, a change in the slope of this line will occur if the autofluorescence of only one of the populations is modulated, and this will impact the compensation matrix.

The bottom line here is to make sure that whichever cells or particles being used for a negative control or population have the same autofluorescence intensity in every channel as the cells or particles being used for the stained controls.

One common instance where this can be problematic is when generating a control for CD14 staining. CD14 is expressed on monocytes, which can be significantly more autofluorescent than other blood cells.

Therefore, if cells are used for compensation controls (e.g. lysed whole blood), then the autofluorescence intensity of the CD14-positive population will be more intense than the CD14- negative population.

This is a good situation for antibody capture beads, which are manufactured to have very consistent autofluorescence levels among particle types.

3. Match your fluorophores.

Finally, be sure that the fluorophore used to generate a compensation control is exactly the same fluorophore that is used in the experimental sample.

For example, avoid using an Alexa FluorⓇ 488-conjugated antibody to compensate for GFP. Even though similar, Alexa FluorⓇ 488 and GFP are different fluorophores and contain different emission spectra, so a compensation control prepared using Alexa 488 will generate a different compensation matrix than one prepared using GFP. Figure 4 below illustrates this difference in emission spectra between Alexa FluorⓇ 488 and GFP (spectra were generated using the BioLegendⓇ Fluorescence Spectra Analyzer)

This extends further to tandem dyes like PE-Cy5, APC-Cy7, and the Brilliant Violet™ dyes, some of which can be tandem dyes.

Due to differences in conjugation efficiency between the donor and acceptor dyes, different lots of a particular tandem may differ in emission spectra and intensity, so make sure that compensation controls are generated using the same antibody conjugate that is used in the experimental sample. This is another situation in which antibody capture beads can be extremely useful.

Fluorescence compensation is not possible without proper controls, so it is critical to spend the time and effort to generate high-quality controls in the preparation of an experiment.

For a compensation control to be considered “good” or “proper”, each compensation control must be as bright as or brighter than the experimental stain, autofluorescence should be the same for the positive and negative populations used for the compensation calculation in each channel, and the fluorophore used must be the exact fluorophore (i.e. same molecular structure) that is used in the experimental sample.

To learn more about 3 guidelines for setting compensation controls in flow cytometry experiments, and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

At the end of many experiments, the question of statistical analysis rears its ugly head. When it comes up, many researchers freeze not knowing how to proceed, and they muddle through as best they can. With some proper planning and forethought, this doesn’t have to be the case.

To resolve this analysis dilemma, it is important to begin thinking about the statistical analysis during the initial designing of the experiments.

This is where some critical decisions need to be made to ensure that, if there are statistically significant findings to be uncovered, the data will be sufficient to support them.

During initial experiment design, consider the following…

1. Power the flow cytometry experiment properly.

Simply put, the statistical power of an experiment is the likelihood that the experiment will detect an effect if there is one to be measured.

The higher the experiment is powered, the lower the chance of making a Type II (false negative) statistical error.

There are a variety of calculators out there and one of the most useful is Statmate, from GraphPad software. Although those using OSX are out of luck with the package, it is a great program to add to your toolkit.

To use this tool, one needs to determine the statistical test to be performed, the threshold (α), and the standard deviation. With this information, a table such as the one shown below is generated.

Figure 1: Output from Statmate.

Across the top are columns for different power, and down the side are the number of samples per group.

The numbers in blue represent the difference between the means between the control and experimental sample.

To use the chart, calculate the difference between the experimental and control means, and consult the blue numbers under the column of the power for the experiment.

Cross-referencing the blue number under the power to the number of samples per group will give you the number of samples needed to be run for the appropriate power.

Setting the power at the beginning of the experiment is the best practice and will provide the researcher with confidence on the possibility of finding a statistically significant result if there is one to be discovered.

2. Establish the threshold (significance level).

The threshold, denoted as the α, is the probability level below which the null hypothesis will be rejected.

This value is historically set at 0.05%.

This value is also related to the possibility of making a Type I (false positive) statistical error and is based on work by Roland Fisher from the 1920s, which he suggested was a “convenient cut-off to reject the null hypothesis.”

Consider the normal distribution, as shown below. If a two-tailed T-test is performed on the data, with a threshold of 0.05%, this is distributed evenly above and below the mean. The white areas represent that 5%.

Figure 2: From https://en.wikipedia.org/wiki/File:NormalDist1.96.png used under the GNU Free Documentation License.

The P value will be compared to the α to determine if the the null hypothesis can be rejected or not.

Since the α is a measure of committing a Type I error, the consequences of a false positive must be considered when establishing the threshold.

A higher threshold makes it easier to find significance, but increases the possibility of the Type I error.

Lower the threshold and it decreases the possibility of a Type I error. Thus, setting the threshold should be considered based on the specific conditions of the test.

Setting the threshold at the beginning of the experiment is a best practice, as it helps establish the probability of committing a Type I error.

3. Clearly state the hypothesis.

At the beginning of the experimental planning, it is critical to understand what the hypothesis being tested is.

If the hypothesis is poorly stated, the rest of the statistical analysis will be inaccurate, or as it is said ‘Garbage In, Garbage Out.’

Since the hypothesis will be used to establish the null hypothesis (H₀), this becomes the most important step in the process, as it forms the basis of why the experiments are being performed.

For example, when asked to determine if a new drug, Pescaline D, increases the number of CD4+ T-cells in patients with Bowden’s malady, one can design an appropriate experiment.

The null hypothesis for this experiment could be stated as: Pescaline D causes no change to the percent of CD4+ cells in patients suffering from Bowden’s Malady.

Setting the null hypothesis at the beginning of the experiment will assist in the design of the experiment, help evaluate the best controls to use, and guide the direction of the statistical test.

4. Choose the correct test.

The statistical test should be identified at the beginning of the experiment.

Based on the null hypothesis, the correct testing method should be clear.

Some common statistical tests, and when they should be used, are listed here:

Figure 3: Suggested statistical testing. A more complete list can be found here: http://www.graphpad.com/support/faqid/1790/

This choice will influence the data that is extracted from the primary analysis.

In the case of the example above, the data would be the percentage of CD4+ cells between the control (untreated) and experimental (treated).

This would be tested using an unpaired T-Test. However, if the experimental design was to take a sample before treatment (control) and treat the patient (experimental), one would perform a paired T-Test.

For those performing T-tests, another consideration is whether to do the test as a ‘one-tailed’ or ‘two-tailed’ T-test.

This is another consideration that has to be made before the experiments are performed.

If the expected change is in one direction — that is, there will be an increase or a decrease — then a one-tailed T-test is appropriate.

On the other hand, if it is not known which direction the change will occur, a two-tailed T-test is the best test to choose. This is defined at the beginning of the experiment to avoid the desire to look at the data and choose a one- or two- tailed T-test at the end.

Choosing the appropriate statistical test at the beginning of the experimental design process is the best practice to prevent bias.

This will ensure that there is no experimenter bias introduced after the data is collected and will also ensure the correct data is extracted from the primary analysis.

5. Know how to plot your data and do it first.

Although it may sound strange, it is very valuable to plot your data before you move forward with your statistical analysis.

Back in 1973, the statistician Francis Anscombe published the now famous Anscombe’s quartet:

Figure 4: Anscombe’s quartet. From: https://commons.wikimedia.org/wiki/File:Anscombe%27s_quartet_3.svg used under GNU General Public License.

These four datasets are statistically identical: including the mean, the sample variance, the regression line, and the correlation coefficient.

Anscombe published this dataset when many researchers were starting to use computers for their statistical analysis, and entering data without graphing it. This dataset was designed to point out the fact that graphing data is a critical first step and an important check on the researcher as well. If something looks odd, it may be odd.

When plotting data, it is good practice to use a plot that shows all the data points.

Take a look at these two graphs.

Figure 5: Percentage of CD4+ cells before (U) and after (T) treatment.

The bar graph on the left shows the same data as the scatter graph on the right.

The difference is that with the scatter graph, it is possible to see that there are two different levels of response in the data after treatment, which is lost in the bar graph. Thus, some data is hidden or not fully evaluated with the bar graph that is visualized with the scatter graph.

Knowing the best graph to use is an important way to convey the important information and supports the statistical analysis that is performed on the data.

It is critical to prepare for your statistical analysis at the beginning of the experimental design process. This will ensure the correct data is extracted, the proper test applied, and that sufficient replicates are obtained so that if an effect is to be found, it will be found. Don’t rely on some magic number of events or samples to determine your experimental design. Rather, rely on the best statistical methods and comparisons to appropriate controls to ensure your data stands up to review.

To learn more about 5 Steps For Accurate Flow Cytometry Statistical Analysis Results, and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

At the most basic level, a flow cytometer is photon counting device. It captures the emitted photons from fluorochromes present on targets — be they cells, beads, or other particles.

These fluorochromes can be attached to antibodies or proteins (like Annexin), free molecules that become fluorescent when bound to a target (DNA dyes), or have different fluorescent characteristics under different biological conditions (Indo-1, JC-1).

Fluorescent molecules are the tools of the trade in flow cytometry and, with continued advances in chemistries, it is helpful to step back and review their essential properties.

Why Understanding The Jablonski Diagram Is Important

When a molecule absorbs a photon of light, an electron is promoted to a higher energy state. This excited state is not stable, and the molecule releases this energy in several different ways.

When the release of energy is an emission of light, it is termed fluorescence.

This process can be visualized by the Jablonski diagram, which is a useful way to model the electronic states of a compound. Such a diagram is shown below (see Figure 1 for a generalized Jablonski diagram of fluorescence).

The emitted photon is of a higher wavelength (and therefore lower energy) than the exciting photon.

This process can be visualized using one of several spectral viewers available from various vendors.

Spectral viewers allow the researcher to understand the emission and excitation properties of a given fluorochrome and help them determine the best excitation line and emission filter set to use for that fluorochrome.

Figure 1. A generalized Jablonski diagram of fluorescence

What Are The 5 Classes Of Fluorochromes?

In general, fluorochromes can be divided into 5 broad categories, which are discussed below.

1. Fluorescent Proteins.

Fluorescent proteins can be categorized into two groups.

The first group is made of natural proteins that encode their fluorescence in the protein structure. These fluorescent proteins include green fluorescent protein (GFP), YFP, RFP, and various derivatives that have been mutated and validated.

Many fluorescent proteins are commercially available in expression vectors to allow researchers to clone them into a vector expressing their favorite targets and follow the expression of the target by using the fluorescent surrogate.

A second class of fluorescent proteins are those derived from the phycobiliproteins found in algae and plants. These proteins use phycobilin cofactors to harvest light and include phycoerythrin (PE), allophycocyanin (APC), and peridinin chlorophyll (PerCP).

Phycobiliproteins, especially PE, are among the brightest fluorescent molecules currently available to researchers (see Figure 2 for the structure of phycoerythrin and its cofactor).

It should be noted that, while very good for flow cytometry, these molecules are not well-suited for fluorescence microscopy because they rapidly photobleach.

Figure 2: (A) the chromophore in phycoerythrin. (B) the structure of phycoerythrin.

2. Synthetic Small Molecules.

Synthetic molecules are a broad class of relatively small fluorescent compounds that have a long history in flow cytometry.

These molecules all contain at least one conjugated double bond system — in rings, chains, or a combination of both —- that is perturbed upon excitation (see Figure 3 for example structures of FITC and Cy5).

Figure 3: The structure of fluorescein isothiocyanate (FITC), from the wiki entry and the structure of Cy5 from the Heterocyclist.

Synthetic dyes are available across the spectrum and come in a variety of configurations that influence solubility and cell permeability.

They are also amenable to chemical modifications for fine-tune targeting through conjugation to lipids, antibodies, and other biomolecules or ligands.

3. Quantum Dots.

In the late 2000s, the use of quantum dots (QDots) became popular.

The QDot is a semiconductor that can be tuned to different emission wavelengths based on the size of the particle (see Figure 4 for an illustration of QDot fluorescence).

These QDots have excellent quantum yields and are very photostable.

The major challenges have been achieving aqueous solubility and creating a surface amenable to conjugation chemistries.

In flow cytometry, QDots are typically excited by the violet laser, although they can be excited by any light below the emission maximum, and so need careful planning when used in panels.

Figure 4. Quantum dye fluorescence is related to particle size. Image from Printed Electronics Now

4. Polymer Dyes.

More recently, polymer dyes have become popular.

Developed by Sirigen and first introduced to flow cytometry by BioLegend, these compounds are extremely bright (BV421 is at least twice as bright as PE) and take advantage of the violet laser, like the QDots.

Since the introduction of Brilliant Violet, additional polymers for UV excitation (Brilliant UV) and blue excitation (Brilliant Blue) have been developed.

The brilliant series of dyes — based on Nobel Prize winning chemistry — work by having multiple segments that absorb light, and when that happens, electrons migrate along the chain.

The polymer backbone itself is fluorescent (such as BV241), or can be coupled to an acceptor fluorochrome, forming a tandem dye (see Figure 5 for an illustration of polymer dye structure).

One upside of these dyes is as appropriate polymers are developed, additional spectral space will become open. In addition to the Brilliant Violet dyes, there are already Brilliant UV (6 dyes) and Brilliant Blue (2 dyes) on the market.

Figure 5: Brilliant Violet dyes. (A) Principles of how the Brilliant Violet dyes work . (B) some of the BV dyes available for flow cytometry. Figures from Sirigen website.

5. Tandem Dyes.

Finally, there are the tandem dyes.

Tandem dyes are a special class of fluorescent molecule that take advantage of Förster resonance energy transfer (termed FRET, or fluorescence resonance energy transfer).

In this process, two fluorochromes are placed in close proximity. The emission of the donor molecule must overlap with the excitation of the acceptor molecule.

When the donor molecule is excited, an electron is promoted to a higher energy state and transfers the energy to the acceptor molecule, which returns to ground state, giving off a photon of light (see Figure 6 for a Jablonski diagram of tandem dye fluorescence).

Tandem dyes are very useful as they can extend the usable spectrum for a given excitation laser.

Figure 6: Jablonski diagram of FRET. By Alex M Mooney – Own work, CC BY-SA 3.0, https://commons.wikimedia.org/w/index.php?curid=23197114

FRET efficiency falls off at the 6th power of the distance, so these manufactured dyes must be protected from conditions that create free radicals, lest the tandem dye can decouple and fall apart. Some of these conditions include light, changes in temperate, age, and fixatives.

Historically, tandem dyes were named as the donor acceptor combination, so researchers are familiar with dyes like PE-Cy5 and APC-Cy7. With the advent of the brilliant violet dye series, this naming convention was discarded. Some dyes, including BV570, BV605, BV650, BV711, and BV786, are tandem dyes, even though there is no indication of this status in their name.

How To Understand Brightness In Flow Cytometry

Ranking these fluorochromes is essential for the researcher, as brightness plays a critical role in panel design.

Brighter fluorochromes should be reserved for critical markers — be they of low or unknown expression, or on rare events.

While it is possible to describe brightness in terms of quantum yield and efficiency, these values are not easy to relate to in a practical sense. Thus, flow cytometrists often discuss brightness in terms of the ‘staining index’, which was described in a paper by Maecker and co-workers (see Figure 7 for an illustration of staining index).

Figure 7: Measuring SI based on Maecker et al. (2004) Cytometry A 62:169

There are several charts online, with varying selections of fluorochromes. The one below from BioLegend has many of the newer dyes.

Figure 8: Fluorochrome brightness chart. The full chart can be found here.

This data is also integrated into the panel design program from Fluorofinder, which assists in the panel design process, meaning you don’t have to flip back and forth between several screens to determine the relative brightness when designing your polychromatic panel.

With the continued development of new fluorochromes, it is critical to stay up on the trends.

While early adoption of new fluorochromes is exciting, there are times that problems lurk that we do not discover till many years later.

Many researchers using Quantum Dots were surprised when Zarkowsky and co-workers published their paper that detailed how copper ions, at nanomolar levels in some commercial fixatives, quenched QDot fluorescence. How many experiments had to be redone because of this finding?

The polymer dyes are reported to aggregate when more than one is used in staining, so the use of the recommended staining buffer is an important consideration when using them in your panels.

Considerations that must be made when choosing fluorochromes include the brightness of the dyes in question, the instrument configuration, and the staining protocol. Each of these factors will impact the quality of the data because of issues related to spectral spillover, staining, loss of signal because of tandem dye degradation, the ability to get an antibody/fluorochrome into a cell, and more. It takes time and effort to develop and optimize a panel. If one fluorochrome doesn’t work, consider why it may have failed and look for alternatives.

To learn more about Why Understanding Fluorochromes Is Important In Flow Cytometry, and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

Reproducibility is the name of the game in science.

For scientific results to be valuable, they must be reproducible. The current crisis in scientific reproducibility has been well-highlighted, and has spurred the NIH to initiate a reproducibility and rigor initiative for grant applications.

There is significant concern about the quality and consistency of antibodies in scientific research.

Bradbury & Plückthun published a letter in Nature, signed by over 100 researchers, that called for standardizing antibodies by moving away from reliance on traditional monoclonal and polyclonal production towards generation of recombinant antibodies.

They estimate that $350 million a year is wasted on bad antibodies in the US alone. They go on to estimate that to produce recombinant binding reagents to 20,000 human genes would cost about $1 billion — three years worth of wasted antibodies.

Until such robust reagents are available, it is the responsibility of the researcher to be vigilant in their use of antibodies.

This is especially critical in the realm of flow cytometry. Until we have validated recombinant antibodies, there are a few steps that can be integrated into the research workflow to ensure that the traditionally produced reagents are working as intended.

1. Always titrate.

The FAb fragment of an antibody binds with high affinity to the epitope against which the antibody was raised. However, at high concentrations, the intended targets become saturated and the antibody will bind to low affinity targets.

This increases the background fluorescence measured by the flow cytometer.

To avoid this off-target binding, it is paramount to titrate antibodies before using them in experimental work.

Titration should be carried out under the same conditions that the experiments will be performed, with the only variable being the concentration of antibody.

Starting with the vendor-recommended concentration, I recommend an 8-point serial-dilution curve for titration of reagents that have not been used previously. It is good to include at least one point near the recommended concentration.

Interpreting titration data can be performed in several ways.

Fig. 1: Histogram view of titration data showing positive and negative populations.

The first approach is to display histograms and estimate the best separation between the positives and negatives by eye (see Figure 1 for an example). This can be useful for well-separated populations, but it is not always clear for dim or rare populations.

To best examine the data, it is recommended that one extract the median fluorescent intensities and use a standardized metric for comparison.

The metric can be fold over background (positive MFI/negative MFI), the Staining Index, or the Separation Index. Each have their strengths, and are illustrated below.

Fig. 2: Concatenated data (left) and plot of three different standardized metrics (right) for sample titration data.

As can be seen in Figure 2, each of these three methods results in the same conclusion for the ideal concentration.

The Staining Index and Separation Index take into account the spread of the data of the negative population, which can be seen in the concatenated file above. This spread represents non-specific binding of the reagent to the negative cells at high concentration.

No matter which way you choose to analyze the titration, knowing the best concentration at which to use your antibody is essential for generating high-quality data.

2. Validate specificity.

The specificity of each reagent used should be validated. While it is assumed that the vendor is doing this, knowing how to validate the reagent in your system is important.

In the excellent article on controls by Hulspas and co-workers, even the same clone from different vendors can perform differently. As shown below in Figure 3, two PE labeled antibodies against human CD34 (clone 581) from different vendors exhibited different characteristics.

Fig. 3: Figure 5 from Hulspas et al. (2009) Cytometry B 76:355-64. True staining is shown by the red circle on the bottom bivariant plot. In both the bivariants and the histograms, there is significant non-target binding from one vendor’s product.

Inspecting cells by microscopy after labeling, or developing positive and negative target lines using techniques like CRISPR and siRNA, can further assist in providing the researcher with confidence in the specificity of antibody binding.

In addition to running the samples required to evaluate binding properties, it is also critical to have the correct controls for interpreting the results.

3. Be wary of isotypes.

Isotype controls continue to be used as a means to “identify” the background binding of a given antibody isotype.

Though this topic has been covered in detail, it bears repeating that in the effort to improve reproducibility, relying on the isotype control to determine background binding is not a robust method.

Andersen and co-workers recently published an excellent paper regarding reducing Fc receptor mediated binding on monocytes and macrophages while also showing the issues related to reliance on isotype controls (see Figure 4).

Figure 4: In the absence of blocking, the isotype control (IgG1) binds at a much higher level than the specific binding for Tie2, a protein known to be expressed at low levels on monocytes. From Andersen et al. (2016) Cytometry A 89:1001-1009.

Maybe it is, as Keeney and co-authors put it, “time to let go” of the isotype control.

4. Integrate critical controls.

Using every control possible — especially during panel development — is essential to establishing that protocols and reagents are working well.

Fluorescent labeling can impact antibody binding and render a good antibody bad. Likewise, some cells can bind the fluorochrome directly.

The use of the isoclonic control can help ensure that binding is specific, and not due to interactions with the fluorochrome.

To use this control, incubate cells with fluorescent conjugate and increasing amounts of unconjugated antibody. Specific binding is established when a decrease in fluorescence is observed (see Figure 5 for an example of an isoclonic control experiment).

Figure 5: As increasing amounts of unlabeled antibody are added to the staining mixture, the intensity of the CD19 population decreases, indicating that antibody binding is mediated through the antibody, and not due to the fluorochrome.

If fluorescence is not reduced in the presence of unconjugated antibody, a different fluorochrome should be considered.

Another useful control is the internal negative control.

Internal negative control cells are usually identified during the panel development and validation phase, and represent a population of cells that are known to not express the target protein.

Thus, a single stained sample can reveal non-specific binding issues with the target antibody, based on the biology of the cell populations within it.

A lot has been written about using the Fluorescence Minus One (FMO) control to identify true negatives based on the spectral spillover into the channel of interest (see Figure 6 for an example).

The FMO control ensures that contribution to the spread of the fluorescence signal in the target channel by the other fluorochromes is accounted for during analysis.

Figure 6: Using the unstained control to determine positivity, as shown by the red line, would cause all cells to appear PE positive. Using the blue FMO line to bound the data, it is clear the cells in question are not PE positive.

Until we have access to well-validated recombinant antibodies produced under tightly regulated conditions, researchers need to exercise good judgment regarding these critical biological reagents. These 4 steps outlined above will help ensure that your results are consistent and reproducible. This will both reassure your reviewers that your data is of high quality, and allow for researchers at other institutions to successfully replicate your results. In addition, identifying antibody duds early on will save you time and money in the long run. Don’t shirk the work of ensuring your antibodies are working correctly and targeting the right proteins.

To learn more about 4 Steps To Validate Flow Cytometry Antibodies And Improve Reproducibility, and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

After running a flow cytometry experiment and collecting the data, the first and only thing any scientist wants to do is analyze the data.

To do this properly in the field of flow cytometry, a quality data analysis software program is required.

The problem is, most of the current data analysis programs are archaic, clunky, or overly complex. The package or packages that were useful in the 1990s have not been updated properly and, as a result, are buggy. Worse, they often require a great deal of training to use correctly.

However, the biggest problem overall with the majority of the data analysis software programs available is their outdated workflow.

For example, most of these programs require you to follow this sequence: load your data into the software program, perform your gating and analysis in the software program, and then spend an immense amount of time exporting your data to PowerPoint, Excel, Prism, and so on, and then further analyzing your data.

…and further analyzing your data, and further analyzing your data, and further analyzing your data…

The Complexity Of Most Data Analysis Software Packages

Why does analyzing flow cytometry data need to be this complex?

Why is there so much busy work involved?

Think about it – in most of these flow cytometry data analysis packages, you first have to do your analysis in that software program, then, as mentioned, you have to export to a different software program like Excel, and then you have to do more data analysis in that software program (i.e. applying Excel shortcuts and equations), then you have to generate your bar charts, pie charts, and whatever else you need, in Excel too.

Wait, there’s more… then you have to copy and paste your charts into yet another software program, such as PowerPoint or Illustrator, not to mention the back and forth you have to do in order to move the raw data from one software program to another. It’s a tiresome and needless process.

At the end of all this, you have your original flow cytometry data files open, as well as dozens and dozens of individual flow plots (at least!), workspace windows, tables, layouts, Excel and PowerPoint files, Prism, .CSV files, and more.

All of these items remain open on your computer because you have to keep them open in order to go back and forth to adjust your analysis.

There’s no reason for it.

What Scientists Want From A Data Analysis Software Program

Data analysis is about producing a final result.

The final result, in this case, is the Excel spreadsheet, or the table, graph, or overall figure you need for your presentation, paper, or grant.

The final result is not 250 flow plots stuck in a workspace.

As such, the ideal software analysis program would allow you to analyze all of your data in a single program. In other words, the ideal software program would allow you to open up an Excel-like table and a PowerPoint-like layout directly in the software program. It would allow you to utilize all of the Excel shortcuts and equations directly in the program itself.

The ideal software program would make it so you didn’t have to constantly export your data to other software programs, or keep dozens and dozens of plots open just in case you need to slightly modify a gate.

Why can’t you just modify your gates within a PowerPoint-like layout and have all of the downstream gates, plots, tables, T-tests, P-tests, numbers, and analyses automatically updated in an instant?

Most importantly, the ideal software program would be simple to use. Instead of requiring complex training, it would be streamlined and as easy to use as Excel or PowerPoint.

3 Advantages Of The FCS Express 6 Data Analysis Program

If you’re using one of the more outdated and complex data analysis software packages on the market, you might want to think about making a change.

This is usually as simple as asking your Principal Investigator or Institution Administrators to update their current software license to something more modern (and more affordable) like FCS Express 6.

If they ask you to justify the change, simply send them the following list…

1. FCS Express 6 is the most user-friendly data analysis program on the market.

If you can use Microsoft Excel or Microsoft PowerPoint, you can use FCS Express 6…

The world of flow cytometry data analysis has changed immensely over the years. It wasn’t that long ago that only core managers, with years and years of training, were the only ones able to analyze flow cytometry data correctly.

This is why most of the data analysis software packages on the market are overly complex. It’s because these packages were created for the top 1% of core managers, not for end users and most core managers.

FCS Express 6 is different. I’ve found this program is ideal for not only core managers, but for end users as well.

The best thing about this program is that anyone, even the most novice flow user, can walk up to it and start creating plots in seconds. This just isn’t the case with other software programs on the market.

I’ve seen even the most novice user sit down, open up an FCS Express 6 PowerPoint-like layout, drag in a data file, watch the plot automatically open up, and then very quickly start doing some very advanced analyses.

2. The FCS Express 6 workflow is modern and aligned with current best practices.

As mentioned above, the workflow of most data analysis software programs is overall complex at best, and flat-out wonky at worst.

Why should I have to load all of my files into some workspace, open dozens and dozens (sometimes hundreds!) of individual plot windows to draw my gates, then open the program’s supplementary tables and layouts, and then export to Excel and PowerPoint, and then go back and forth for hours and hours (or even days) tweaking each individual gate in each individual plot, while simultaneously updating my exported files with every single tweak?

It’s exhausting.

The larger problem here is that this outdated and complicated workflow can result in data analysis errors and experimental irreproducibility. It’s just not feasible to expect new flow cytometry users, or any user for that matter, to mentally keep track of every single change in every single individual plot window, every single time a change needs to be made.

With FCS Express 6, however, any user, new or advanced, can simply just open up the program’s PowerPoint-like layout and Excel-like table and start creating plots, tables, charts, T-tests, P-tests, and everything else, right in the program itself… all the way to completion of whatever flow table or figure you need for your paper or grant.

The best part is, every change made to any gate in any plot results in everything downstream being automatically updated.

Each and every plot is already in a single PowerPoint-like layout, not in dozens and dozens of individual plot windows. This means you can see these changes happening across the board in real-time.

3. FCS Express 6 has the most rigorous quality control standards.

There’s a reason that why Express 6 is used in more flow cytometry clinical analyses than any other data analysis package.

FCS Express 6 has more than 100 reference labs using the clinical version, including 3 of the world’s largest laboratory companies (Labcorp, Quest, and Genoptix).

FCS Express 6 also provides an IQ/OQ package for validation. IQ stands for Installation Qualification and OQ stands for Operational Qualification.

If, for example, you are a pharmaceutical company who is acquiring a new piece of equipment, you will need to design specifications that define exactly what’s in that piece of equipment. This is where the IQ/OQ package comes in.

Finally, FCS Express 6 also provides comprehensive 21 CFR Part 11 compliance features. The Food and Drug Administration’s (FDA’s) Electronic Records and Electronic Signatures Rule (“Part 11”) defines the requirements for submitting documentation to the FDA in electronic form, as well as criteria for use of electronic signatures.

FCS Express 6 meets these requirements and is designed to maintain both the machine-readable metadata and human-readable reports.

FCS Express 6 has the following features that can enhance your lab’s Part 11 compliance program: security, access limitations, authority checks, record protection, audit trails, electronic signatures, and much more.

In fact, over 300 independent security permissions can be set for each security group, giving you tremendous flexibility in defining your own access limitations.

Overall, FCS Express is the ideal data analysis software program to use when analyzing your flow cytometry experiments because it is the most user-friendly program available that is both aligned with current data analysis best practices, and maintains rigorous quality control standards.

To learn more about 3 Advantages FCS Express 6 Has Over Other Flow Cytometry Data Analysis Software Programs, and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.

Written by Tim Bushnell, Ph.D

Reproducibility is one of the hallmarks of the Scientific Method. One reason scientists publish their results is so that other labs can attempt to reproduce the results and extend the investigation into new areas.

From the Begley and Ellis commentary in Nature, to the development of the Rigor and Reproducibility initiative at NIH, to articles in the mainstream media, reproducibility is on everyone’s mind. This is not a bad thing, and following the best practices in flow cytometry provides investigators, peer reviewers, and colleagues with more confidence in your data.

There are several areas within the process of developing, implementing, and reporting a flow cytometry experiment where a little additional work, attention to proper controls, and careful planning will ensure reproducible data generation.

Instrumentation: Characterization, Optimization, and Quality Control.

The first line of defense for good, reproducible data is the instrument.

A properly maintained machine ensures that it is not introducing error in the data. A researcher needs to know that when they find a difference, it’s due to the experiment, not the instrument. Thus, quality control is the key.

In the recent “Best Practices” article from the ISAC Shared Resource Laboratory Taskforce (Barksy et al. 2016), the authors discuss the importance of QA and QC, and highlight the different possible levels — including the idea of an external audit.

It is critical to remember that QC is only good if it is being monitored. Running beads without looking back at the trends over a week, a month, or a quarter doesn’t help anyone.

The end-user should not be shy about asking those in charge of the instrument to see the QC reports or Levey-Jennings plots (Figure 1).

The end-user can also take ownership of QC for their experiments by incorporating a bead standard that is run before each experiment in the series, and monitored. Such a practice was discussed in a recent paper by Misra and co-workers (2016).

Figure 1. Levey-Jennings graph of QC data monitoring PMT voltage changes over a 100-day period. The blue line represents the mean, and the black lines +/- 1 and 2 standard deviations around the mean. Depending on the experimental needs, this data can be used to determine when an intervention is necessary before running a critical sample.

Another area that investigators should look for, regarding the instrument, is optimization of the system. At a minimum, the optimal PMT voltages and linear dynamic range should be available to all users. Additional characterization, such as the protocols discussed in this paper by Perfetto and co-workers (2012), can help with panel design.

It may seem like a lot of work, but what is the cost of a ruined experiment? Lost data? It is critical that researchers begin to ask for this level of characterization from their instruments, and that those operating them are given the time and resources necessary to perform such characterizations.

Reagents: Cells, Antibodies, and Buffers.

In many cases, our flow cytometry experiments will be measuring a biological process in cells and cell lines using fluorescently tagged antibodies.

The issue of cross-contamination of cell lines is well-documented, and it is incumbent on the investigator using cell lines to provide validation of the cell line in question. With advances in next generation sequencing and proliferation of commercial services to perform this validation, it is critical that you know the cell line you are using.

A more deeply-rooted problem may be naming conventions. In a recent Nature article, Yu and colleagues suggested a new framework for naming cell lines.

Antibody naming conventions have benefited from the Human Leucocyte Differentiation Antigens (HLDA) workshops, which share information on their website about different antigens and antibody clones available for investigators.

Regarding antibodies, optimization of these reagents is an essential component to good flow cytometry. Titration is the first important step towards characterization of a given antibody. Performing this assay allows for determining the best concentration for your experimental assays, as well as validating that the antibody is working.

Figure 2. Titration of an antibody. On the left is the concatenated file with increasing antibody concentration, and on the right is the calculated Staining Index calculated per Maecker et al. (2004).

Additional validation steps of the antibody is a critical discussion that is ongoing in the literature.

While we currently rely on vendors to ensure their processes provide us with what the label says, a recent letter in Nature by Bradbury and Plückthun (and 110 co-signers) highlighted our dependence on protein binding reagents and their current limitations, as well as the magnitude of research dollars lost to poor quality reagents. It will bear watching to see how the industry moves in the directions that this letter advocates.

A quick comment on buffers and reagents — many investigators rely on purchasing pre-mixed reagents as a way to improve consistency and lot-to-lot variation over house-made reagents. It is important to ensure that you have a proper training and validation protocol for reagents, regardless of their origin, to make sure that they are not negatively affecting your data.

A cautionary tale on using pre-made reagents is found in this paper by Zarkowsky and co-workers (2011). Here, the researchers noticed a loss of quantum dot fluorescence with some lots of commercial formalin solutions.

After characterization of different lots, it was discovered that copper ions, a micro-contaminant, were the culprits (Figure 3). The solution to the problem was simply to add EDTA to the fixative, which chelated the Cu++ ions.

Figure 3. Figure 1 from Zarkowsky’s paper, showing the effect on Qdot fluorescence of three different lots of fixation buffer (+/- EDTA).

One can only imagine how many investigators lost data between the adoption of quantum dots in flow cytometry and the publication of this paper. It serves as a cautionary tale for researchers adopting newly available fluorochromes to delve a little deeper into their buffers in the case of anomalous results.

Process: Monitoring Human Error.

Variability in data can be a result of sample processing. There are many steps between isolating the cells and acquiring data. Development and use of standard operating procedures (SOPs) can help mitigate variations. A mainstay of clinical labs and GMP production facilities, more and more basic research labs are beginning to adopt SOPs. The aforementioned Barksy et al. 2016 paper spends a good deal of time talking about their role in flow cytometry.

Another way to monitor the sample preparation process is to develop and validate a “reference” or “process” control. This is a standard sample that is run each time the assay is performed.

Knowing how this standard sample performs in the assay will allow the researcher to set acceptable limits and know if there was an issue with part of the process (Figure 4). The severity of the deviation determines whether it is possible to troubleshoot and move the assay forward, or if the run has to be performed again. Likewise, it can point to instrument issues that developed between the QC and the time the investigator ran the assay. Of course, as with any QC system, it must be monitored.

Figure 4. Process control sample showing the range of expression for different populations in an assay.

Analysis: Extracting the Correct Data.

The fourth area for consideration to improve reproducibility of flow cytometry results is data analysis. This is the area where the ‘science’ of cytometry meets the ‘art’ of analysis, and the one that causes the most headaches.

At present, most flow cytometry analysis relies on manual bi-variant hierarchical gating strategies. As more robust automated workflows are developed, they may become the dominant analysis methods, reducing or removing human bias from the process. Until that time, however, a few simple rules can help improve analysis and reduce errors.

Before analysis begins, it is important to understand what the biological question is, what data needs to be extracted from the experiment, and what downstream analysis will be performed on the data. This will drive the development of the analysis strategy.

To develop a robust analysis strategy, one must understand best practices for gate setting. This is the basis of a separate blog post, as there are many factors involved in that process.

Good analysis requires using appropriate controls, understanding display options, and developing consistent rules, once again leading to the SOP.

In the recent update to FCS Express, they included an SOP builder in the software. This allows for the creation of an analysis SOPs for lab members to follow. Checkpoints can be built in throughout the analysis, so that the researcher moves through step-by-step, and will not proceed to the next step unless the requirements are met. It’s a great feature for a lab performing a standard analysis over time, so that everyone performs the analysis the same way.

Consistent analysis is critical, as was shown in a study published by Maecker and coworkers (2010), which compared the analysis of prestained cells in 15 experienced labs versus analysis in one central lab. The results showed a mean CV of 20.5% across the four samples for analysis by the remote labs, versus 4% when the analysis was performed in a central lab. That level of data spread can make it extremely difficult to find rare differences, thus reducing the likelihood of advancing the field.

For consistency in gating, relying on more normal shapes for gates (rather than free-form drawings) is an important consideration (Figure 5). Couple that with the use of FMO controls, establishing cut-off percentages for gates, and acquiring appropriate biological controls, one can improve the consistency of results.

Figure 5. Sample analysis showing a gating strategy for analysis of a multi-year study.

Finally, because there is a great deal of information that cannot makes its way into a standard materials and methods section, the ISAC Data Standards Task Force developed and published the “Minimal Information about a Flow Cytometry Experiment” or MIFLowCyt standard. This standard establishes a detailed checklist covering four major areas:

• Experimental overview
• Sample and specimen description
• Instrument details
• Data analysis details

Papers published that are compliant with this standard ensure an additional level of information is available to researchers seeking to reproduce the data.

At present, this standard has not been widely adopted, but is strongly encouraged for submissions to Cytometry A. A checklist is available from the journal to assist in ensuring a submitted paper has met the reporting requirements of this standard.

Additionally, ISAC has developed and supports the FlowRepository, a database where investigators can submit their data. This database, first described in a paper by Spidlen and coworkers, allows peer reviewers access to the raw data used in a paper so they can examine the data to improve the review process.

After publication, the data is available to investigators as a way for them to compare their results with published data.

An excellent example of the FlowRepository in action are the Optimized Multicolor Immunofluorescence Panels (OMIPS). These papers describe the development of new polychromatic flow cytometry panels for the research community, and are required to show the gating strategy used by the authors. Having the primary data available to the research community again offers a way to combat issues in reproducibility.

In conclusion, there are several areas that researchers can focus on to improve the reproducibility of their flow cytometry experiments. From instrument quality control, through validation of reagents, to reporting out the findings, a little effort will go a long way to ensure that flow cytometry data is robust, reproducible, and accurately reported to the greater scientific community. Initiatives by ISAC have further offered additional levels of standards to support these initiatives, which were developed even before the Reproducibility Crisis came to a head in both scientific and popular literature.

As scientists, we owe it to ourselves, our colleagues, and the public to ensure the data we generate is of the highest quality. It is Isaac Newton who is attributed to saying, “If I have seen further, it is by standing on the shoulders of giants.” Together, following the steps outlined above, we can each stand on the shoulders of our colleagues to move scientific discovery forward, with the ultimate goal of improving the health and well-being of our fellow man.

To learn more about 4 Ways To Achieve Reproducible Flow Cytometry Results, and to get access to all of our advanced materials including 20 training videos, presentations, workbooks, and private group membership, get on the Flow Cytometry Mastery Class wait list.