Interpreting and Exporting Data from Cytobank

[Hende433]

Hello everyone,

I am new to mass cytometry and do not have any background in flow cytometry. I have a couple of questions concerning some data I recently obtained from a myoblast cell line. I do all of my data analysis in Cytobank.

First of all, I generally see a very large spike at the extreme low end of the x-axis on my X Histograms. I do gate for cells using the Ir intercalator, but I wonder if I am not doing this properly? Does anyone else see this or know what it is?


Also, I am wondering if there is a way to export the data (not the statistics) so that I may find the intensity values for the highest and lowest points of the X histogram. Does anyone know of a way to do this?


Any insight is very appreciated.

Thanks!

[mleipold]

The histograms you show are a measure of the signal intensity in that channel. Each cell event's value is graphed at the appropriate part of the distribution. The height of the curve at each point corresponds to the fraction of the population at that value.

The large spike at the extreme left/low end are all the zeros/low values. In fluorescence flow, you get a distribution of values at or below zero due to compensation and autofluorescence. Since CyTOF essentially doesn't have "autofluorescence"/biological background (except for Iodine-127), and there is no compensation due to spillover from one channel to the next, the cell events with a "0" value in that channel are measured as "0".

Graphically, the "0" values get distributed between -1 and 0, so that the "0" peak isn't quite as tall as it would be (and therefore your other peaks are proportionately higher).


In other words, the left spike *is* true.


Mike

[ChrisCicc]

Hi Michelle,

I'll try to answer your questions in context of Cytobank functionality.

I do gate for cells using the Ir intercalator, but I wonder if I am not doing this properly? Does anyone else see this or know what it is?

If events are appearing in your histogram that you believe should be gated out, my assumption would be that the active population isn't set correctly. As a note, in Cytobank, the term population is generally synonymous with what is often called a "gate". Take for example this gate that I've drawn on a histogram in the gating interface:



If the working illustration is set up how it is in the example below, I wouldn't see the effect of that gate on my plots. Notice that "Ungated" is the selection in the populations figure dimension. As for the red arrows, they are pointing out how setting an axis to "Use Panel/Channel Values" in conjunction with the channels figure dimension can be used to create figures:

http://i.imgur.com/MxzLJW1.png

In order to filter the events that make it to the plots that are displayed in the illustration, the "Range" population/gate needs to be selected. In the example below, I have simply changed the selection in the populations figure dimension and now the histograms are displaying only the events that fall into the Range gate. The histogram that the gate was originally drawn on (191 channel) reflects this clearly:

http://i.imgur.com/5d3Xvtd.png

Also, I am wondering if there is a way to export the data (not the statistics) so that I may find the intensity values for the highest and lowest points of the X histogram. Does anyone know of a way to do this?

No problem. You can elect to export data in FCS or text format. If you elect text, what you will end up with is a spreadsheet that gives you a list of events and their intensities on all the channels in the data set. To do this, simply click the link for "export events" to the left of the working illustration figure dimension boxes. Before you do this, however, you will want to set up your illustration to affect your export, such as in the image below. An individual spreadsheet will be created for each combination of population and FCS file. In the example below, I would get three files. If I had selected an additional population, I would get six files (two populations times three files). The exported data will be gated according to the selected populations.



After clicking the link there will be different options to configure for the export. The one that may require more consideration is "Select transformation". This is referencing the scale settings you have in this Cytobank experiment. If you choose to export without transformation, the data will be written with raw values according to what the machine originally exported. Applying transformation will rewrite the data values according to your scale settings. Data is visualized with a transform applied, so this option may be best if seeking to preserve the relationships you see in plots in the exported files.



Below are small snapshots of the same export. One has been transformed (low values) and the other has not (higher, raw values).





Hope this helps!

Chris