February 28, 2013  |  Education, Uncategorized  |  By  |  0 Comments

My First CyTOF Experiment

Having recently joined Cytobank and with little practical experience with the CyTOF, I headed over to the Nolan lab to do my very first CyTOF experiment. Every year, the Nolan lab hosts a phospho-flow course where a group of interested researchers fly to Stanford to learn how to perform a phospho-flow experiment. During my day there, I used a protocol in development for this course to generate data on the CyTOF and subsequently used  Premium Cytobank for analysis. This is a record of my experiences that will hopefully be helpful to those of you just starting out in the world of mass cytometry as well.

Experiment Preparation and Data Collection:

The first thing that you’ll notice when starting to plan your panels is that selecting antibodies can be easier than for traditional flow experiments due to the lack of compensation when using the CyTOF. The course protocol has a pre-planned panel, so we didn’t have to titrate or check the availability of reagents; however, both of these actions are necessary for a lab starting with the CyTOF.

  • When selecting mass cytometry antibodies, note the availability, label intensity, and antigen density.
  • Fluidigm Sciences sells many pre-conjugated antibodies and labeling kits for mass cytometry reagents.

The experiment we performed consisted of stimulating PBMC’s with different growth factors and then measuring the levels of p-STAT3, p-STAT5, p-ERK, and p-p38 after stimulation. When we collected our samples, the plasma had already been on for a few hours and the instrument was already QC’d so we were able to walk up and start running. While collecting the data, we annotated the channel names with the antibodies we used in the experiment.

  • Be careful of metal ion contamination from common lab chemicals like detergents and buffers
  • Resuspend cells at 5×10^5 in ddH2O just prior to collection
  • Acquisition often takes longer than a flow cytometry experiment- plan accordingly
  • Annotate channel and sample names at the cytometer- Cytobank will automatically transfer them to the analysis


Once the data were collected and exported from the CyTOF, we created a new experiment on Cytobank and uploaded the files. Since we annotated at the cytometer, all of our channel names transferred automatically so we could proceed directly to gating. The experiment used cultured PBMCs for the experiment, so we gated the CD4+ T cell, CD8+ T cell, and CD33+ Monocyte populations. The next step was to add the stimulants to the Conditions figure dimension in order to compare the different signaling responses.

  • When beginning analysis on Cytobank, take a moment to think about which Figure Dimensions you’ll need to annotate. We had Channels, Populations, and Conditions. (You may have Timepoints, Dosages, etc.)
  • Once you’ve put in the up-front time annotating Figure Dimensions, it’s easy to arrange and rearrange plot layouts, and switch among the many Plot Types available.

After annotating the appropriate figure dimensions, we created illustrations within Cytobank. The figure dimensions (Channels, Populations, and Conditions, in this case) are configurable so that any can serve as the rows, columns, or tables in the illustration, enabling the user to rearrange the figure at will. For our first figure, we decided to use heatmaps because they are an effective way to convey information in a compact space, and Cytobank shows the user the underlying data when mousing over the cell of interest.


The second figure is a histogram overlay to better visualize the change in phospho-protein signal, as it can be difficult to convey some nuances in the data, like a bimodal signal, in a heatmap.


The third figure is the gating strategy, and is included at the end of each figure exported from Cytobank to provide context. Gating strategies are important because it is difficult to derive meaning from a figure without an accurate idea of which cells are included in an analysis. After setting up the figures we wanted to keep, we saved them as illustrations and proceeded to export them as PDFs.


As this was a relatively simple experiment, these three figures are a good starting point for digging into the data. However, Cytobank has many more statistic and plot options available for comparing and visualizing data generated by the CyTOF. Notable among these is SPADE, an algorithm and visualization tool which links groups of similar events and allows overlays of signaling responses to help see the implications of a stimulus across all populations in the experiment. The results of the algorithm still need to be interpreted, similar to gating during analysis, but once the figure is properly annotated it becomes a very powerful tool for quickly visualizing data with many parameters and exploring interactions that may have otherwise gone overlooked.


Having the opportunity to run an experiment on the CyTOF was both informative and encouraging – it was different than a traditional flow cytometry experiment for some of the preparation and sample acquisition, but almost identical during the data analysis— the only difference being the increased number of parameters. As CyTOF data become more prevalent in the literature and at conferences, understanding the data generation and analysis processes will aid in data interpretation. If you have the opportunity to work with CyTOF data, send us your thoughts on what would accelerate your analysis. At Cytobank, we are continually evaluating and implementing analysis and visualization tools, and we value working closely with the cytometry community to ensure that we’re working to meet your needs and enhance the field.