FAQ  •  Register  •  Login

'towers' during acquisition in specific channels.

Forum rules
Please be polite and civil. We know that troubleshooting is vexing...
<<

shaunf

Participant

Posts: 2

Joined: Fri Apr 22, 2016 7:36 am

Post Mon Apr 25, 2016 8:28 am

'towers' during acquisition in specific channels.

Dear All,

A question about a recent experiment and the unusual appearance of staining over time in CD138 -- wonder if anyone has seen similar and tracked down the cause? We stain CD138 using CD138-APC followed by APC-176Yb and found that in contrast to previous experiments, staining over time was non-random, with discrete clusters of positive cells in the first 2-3 tubes, and unexpectedly few positive cells in the remainder. The samples are murine bone marrow/spleen preparations, with number of events acquired in each tube from 80k to 497k (first 4 tubes: 497k, 256k, 104k, 77k)

A link to a .pdf illustrating the problem is as follows:
https://www.dropbox.com/s/oimrenr2nrorr ... 8.pdf?dl=0

The first slide is the intensity of CD138 staining (y-axis) in each of 14 tubes (1-14) run consecutively over time (x-axis). In tubes 1-3 in particular positive cells appear in patchy towers and thereafter hardly any positive cells are seen at all. We use barcoded mixed samples, mixing 4 samples in each tube, so feel it is unlikely that this is related to the staining of an individual tube.

The second slide is CD138 staining from our previous experiment using a very similar staining protocol. This is much more what we expected to see: a much more random distribution of CD138 positive cells over time with positive events in all tubes.

The third slide is another marker CD19 (149Sm) from the same experiment as the unusual staining again we are perhaps losing positive events over time and a few towers are noted in tube 2.

In other words, to summarise, it seems to me that we have far fewer CD138+ events than expected, and that the distribution of positive events over time is non-random. I've searched the forum as best I can and have not been able to find similar; our hypotheses are primarily ?too concentrated a sample or clumping (but filtered immediately prior to running) ?machine-related technical phenomenon. Any thoughts as to what might be causing this, and anything we should be thinking about prior to our next run in a month's time?

Many thanks, in advance,


Shaun.
(Visiting fellow, INSERM, Paris)
<<

AdeebR

Grand master

Posts: 169

Joined: Thu Mar 13, 2014 5:58 pm

Location: NYC

Post Mon Apr 25, 2016 5:24 pm

Re: 'towers' during acquisition in specific channels.

Hi Shaun,

Do each of these plots correspond to a single tube (containing your 4 barcoded samples) or are these individual samples post barcode deconvolution (i.e., are the "spikes" distributed consistently throughout all the barcoded samples in a single tube or are they appearing in a specific sample)?

What samples were acquired prior to acquiring these samples? These spikes look to me like potential sample carryover from previous samples. We've seen some instances where we've seen positive spikes in a sample that shouldn't contain cells positive for that parameter, and have determined these are carryover from the previous sample that did contain positive cells in that channel. I suspect the spike pattern is because the cells have stuck to the tubing and didn't get washed out completely and are released in bursts.

Adeeb
Adeeb Rahman
Icahn School of Medicine at Mount Sinai, NYC
<<

shaunf

Participant

Posts: 2

Joined: Fri Apr 22, 2016 7:36 am

Post Tue Apr 26, 2016 9:09 am

Re: 'towers' during acquisition in specific channels.

Dear Adeeb,

Thanks for the quick response; the plots I've posted represent actual tubes acquired (i.e. mixed barcoded samples). I've normalised the resulting .fcs files (using DVS beads) but they had not been debarcoded. The following link is a CD138 / time plot for the four debarcoded samples; these are the four samples that were mixed to make up the first two tubes -- you'll see that they are quite similar. Note that they therefore represent events from a number of tubes.
https://www.dropbox.com/s/w963o4tgrriau ... 5.pdf?dl=0

Ours was the first experiment run in the day; prior to this was the standard CyTOF QC panel. If this phenomenon represents washed over cells, would corresponding spikes be expected in other channels? CD138+ cells should be mostly CXCR4 positive, but we don't see corresponding spikes in that channel:
https://www.dropbox.com/s/jpnkwvcp35dc4 ... 4.pdf?dl=0

Best regards,

Shaun.
<<

AdeebR

Grand master

Posts: 169

Joined: Thu Mar 13, 2014 5:58 pm

Location: NYC

Post Tue Apr 26, 2016 4:46 pm

Re: 'towers' during acquisition in specific channels.

Carryover does seem unlikely if your samples were the first of the day. However, I do want to note that in the carryover scenario I described, this could be carryover of cells stained with a different panel (i.e., not CD138+ cells per se, but cells that were positive for a different antibody stained in the same channel) so the expected relationships between parameters in a given panel may not apply.
Adeeb Rahman
Icahn School of Medicine at Mount Sinai, NYC
<<

GregBehbehani

Master

Posts: 85

Joined: Tue Apr 12, 2016 10:17 pm

Location: The Ohio State University, Columbus, Ohio

Post Tue Apr 26, 2016 8:38 pm

Re: 'towers' during acquisition in specific channels.

Hi Shaun,

I would agree with Adeeb.

In my experience, this type of pattern (bursts of unusual cells) is carryover or clumping until proven otherwise. Was this run on a CyTOF 1 or CyTOF 2 (with valves and a syringe pump)?

If you're not seeing positivity on CXCR4 as expected, I would be even more concerned. I assume you've checked to see that the cells in these spikes of events are indeed barcoded, right? Are you using the Fluidigm Pd barcoding, or some other system?

I would try to find out what samples were run previously, if they were similarly barcoded, and if what you're seeing could represent carry-over from one of those samples. The "stickiness" of cell samples can vary greatly. Alternatively, your cells of interest could be forming clumps in a cell-type specific way (possibly due to the ant-APC secondary), if this happens after barcoding, the resulting clumps would be barcoded. It's also important to know how the machine is cleaned and QC'ed each day. Is the cleaning sufficient to strip remove all metal from the system? Were beads run long enough, with other non-bead channels recorded, to know if there was any carryover happening prior to the start of your run?

Greg

Return to CyTOF troubleshooting

Who is online

Users browsing this forum: No registered users and 9 guests