FAQ  •  Register  •  Login

Nd148/Pr141/La139 background issues

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

AdeebR

Grand master

Posts: 169

Joined: Thu Mar 13, 2014 5:58 pm

Location: NYC

Post Tue Mar 18, 2014 7:37 pm

Nd148/Pr141/La139 background issues

Hi CyTOFers,
I've encountered a strange background issue that I'm trying to resolve. I'm occasionally seeing smears of bright events in my La139 channel, which spill over into my Pr141 and Nd148 channels (or vice versa I suppose). I realize that there's not supposed to be any crosstalk between these channels, so I'm pretty confused about what's going on. I've seen this in a couple of different experimental settings varying in severity, but the worst example has been with a lysed/fixed whole blood sample that was stained with CD16-Nd148, Siglec8-Pr141 and FceRIa-La139.

Has anyone else encountered anything like this?

I'd like to attach a PDF with some sample plots showing the issue, but I don't see an option to do this.
Adeeb Rahman
Icahn School of Medicine at Mount Sinai, NYC
<<

mleipold

Guru

Posts: 5792

Joined: Fri Nov 01, 2013 5:30 pm

Location: Stanford HIMC, CA, USA

Post Tue Mar 18, 2014 9:07 pm

Re: Nd148/Pr141/La139 background issues

HI Adeeb,

If you look at the bottom left of the posting space, there are two tabs: Options, and Upload Attachment. You can use that second tab to attach a PDF.

If you can, please show your *entire* gating process, from Ir-Ir on down to Live Intact Singlets (or whatever you then do the marker-based gating on) and then all of your marker-based gating.

I don't think that this is "spillover" as such: those aren't M+1, M-1, or M+16 from La139. In whole-blood, a good Lyse is essential, as are good washes after to remove debris from RBCs, etc. It can also be useful to include RBC and/or platelet markers to "dump", as these often are "sticky" and give cell events that make no biological sense.


Mike
<<

AdeebR

Grand master

Posts: 169

Joined: Thu Mar 13, 2014 5:58 pm

Location: NYC

Post Wed Mar 19, 2014 1:15 am

Re: Nd148/Pr141/La139 background issues

Thanks Mike.

The PDF should hopefully by uploaded.

I agree that this doesn't make sense as "spillover", but it certainly looks like it. I should also note that these are the only 3 channels where theses events appear.

Incidentally there is one other issue that I want to point out on the attachment. You can see that there are some bright Eu151+ events visible below the bead population in the 2nd plot, which is something that I've come across in a number of experiments (sometimes more obvious than this). I've taken to excluding them when I draw my Eu bead exclusion, as I've done here, but it would still be nice to have a better explanation as to what these events are.
Attachments
140224_staining_issues.pdf
(365.44 KiB) Downloaded 453 times
Adeeb Rahman
Icahn School of Medicine at Mount Sinai, NYC
<<

Ofir

Master

Posts: 75

Joined: Thu Nov 07, 2013 12:46 pm

Location: US, CA

Post Wed Mar 19, 2014 5:24 am

Re: Nd148/Pr141/La139 background issues

Hi Adeeb,
Thanks for sharing your data.
I suspect that your events are actually triple positive (139, 141, 148) and the reason is that they are doublets or clumps of cells. We have had similar issues when our samples were run too concentrated. We now inject 5X10^5 cells / ml which seems to get rid of most doublets.
Three more relevant points on gating:
1) Event_length is a linear parameter, and is best displayed and gated on a linear scale. I find that for lymphocytes / PBMCs most of my doublets are >=40. Try gating on Event_length 20-38 and see if you reduce the triple stained population frequency. Also look at the inverse gate to see if that population is enriched there. If so, that is a strong indication that you are having doublet issues.
2) To exclude doublets I use a very strict gate for Ir191/Ir193 such that the upper limit is just at the top of the central population. If you look closely you will find doublets just above that central area of cells. It's a little tricky because this is a log parameter on a linear measurement (DNA content). In your example you did in fact exclude the Ir193 high cells, so that may be fine. Have a second look anyway.
3) Your bead exclusion gate (Eu151/153) also excludes single positive events for Eu151:CD123, which seem real to me. I typically include everything but the beads. In FlowJo you can invert a gate, which can be convenient for this purpose.

I would be happy to know if this helps.

Best,
Ofir
<<

AdeebR

Grand master

Posts: 169

Joined: Thu Mar 13, 2014 5:58 pm

Location: NYC

Post Wed Mar 19, 2014 2:56 pm

Re: Nd148/Pr141/La139 background issues

Hi Ofir,
Thanks for sharing your thoughts.

You are correct in that these events are triple positive for the three markers.

However, I realized that my initial post might have misrepresented the experiment a bit - the sample in this case were stained with a 20-parameter staining panel, it just so happens that these three channels are the only ones where I see this strange signal.

Based on this, I don't think this can be explained by doublets/cell clumps (though I will certainly try a tighter gate to see if it helps), since doublets would presumably appear in multiple channels and not just these three. Also, the events are not positive for other markers (e.g. CD66b) which would be expected if they were true clumps of CD16+ neutrophils.

I should also note that this sample was run at <300 events/s.

Regarding your third point, I agree that this exclusion gate seems like it's excluding single positive events, and the rationale for this exclusion is simply that they don't look right (which admittedly is not the most robust reason). However, I know which cells are supposed to express CD123 in the sample (basophils and pDCs), and some of these events are several logs brighter than the true CD123+ population, which doesn't make biological sense. I've also come across these Eu bright events in multiple panels using different sample types and antibodies on the Eu151 channel, which suggests that these are not a true population. I certainly think that it would be nice to have a better explanation as to what these events are.
Adeeb Rahman
Icahn School of Medicine at Mount Sinai, NYC
<<

mleipold

Guru

Posts: 5792

Joined: Fri Nov 01, 2013 5:30 pm

Location: Stanford HIMC, CA, USA

Post Wed Mar 19, 2014 3:21 pm

Re: Nd148/Pr141/La139 background issues

Hi Adeeb,

I would join Ofir in advising a tighter gate on the Event_length.

I would also be curious to see where your "weird" cells fall in other parameter space. In other words, gate on them, and then look at the other staining dimensions.

Have you tried gating on the small CD45bright (~1e3 counts) CD61- population? I'm not that familiar with platelet staining (most of my work is with PBMC), but as I understand it, that smaller population should be leukocytes/lymphocytes, whereas your larger CD61-mid (1e2) CD45-mid (1e2) population would be platelets and other things. If it *is* an issue of platelet stickiness, then gating on those might get rid of the weird population.

Do you have CD66b or something to help you clean up your populations even more?


Regarding the Eu151 bright population: there was a thread in Troubleshooting or Data Analysis recently about a funny bead lot someone got (I think it was Ian Frank, from Benaroya). If the beads are "leaking" europium, then that might explain it.....are those ultra-bright events also bright for the rest of the bead channels?


-Mike
<<

AdeebR

Grand master

Posts: 169

Joined: Thu Mar 13, 2014 5:58 pm

Location: NYC

Post Thu Mar 20, 2014 11:54 pm

Re: Nd148/Pr141/La139 background issues

Thanks for the responses.

I tried drawing a very tight singlet gate and that unfortunately didn't help the issue. I've attached a modified gating layout, where you can see the Pr141 ultrabright events are primarily of lower event lengths.

Incidentally, I was under the impression that for singlet exclusion on the mass cytometry data you were just supposed to eliminate the Ir high events. Do you typically always restrict event length to the <50?

Mike - the two populations you see are CD45bright lymphocytes and CD45intermediate granulocytes (I do have CD66b in the panel, which I've used to confirm this). Most of the CD61hi platelets are Ir low/negative and are excluded in the first gate. The CD61 expression on the CD45 populations probably represents neutrophil/platelet aggregates.

I tried gating on the lymphocytes and granulocyte populations separately and you can see that both still contain the weird bright events.
Attachments
140224_staining_issues2.pdf
(373.94 KiB) Downloaded 390 times
Adeeb Rahman
Icahn School of Medicine at Mount Sinai, NYC
<<

mleipold

Guru

Posts: 5792

Joined: Fri Nov 01, 2013 5:30 pm

Location: Stanford HIMC, CA, USA

Post Fri Mar 21, 2014 12:10 am

Re: Nd148/Pr141/La139 background issues

Hi Adeeb,

I'm not familiar with CD61, really. However, if I'm understanding your gating strategy, you're calling "lymphocytes" as CD45+ CD61-. Wouldn't that be "leukocytes", rather than lymphocytes? In other words, don't you still need CD14/CD33 to gate out monocytes (++) from your lymphocytes (--)?

I would be curious whether those CD16bri-Siglecbri cells are in *all* your lymphocyte populations (CD4+, CD8+, B cells, NK cells), or whether they're mostly/totally in your monocytes.....in our experience, monocytes tend to have higher background (sticky) than the others.


Mike
<<

AdeebR

Grand master

Posts: 169

Joined: Thu Mar 13, 2014 5:58 pm

Location: NYC

Post Fri Mar 21, 2014 1:33 am

Re: Nd148/Pr141/La139 background issues

Hi Mike,
In my experience, in whole blood samples (particularly ones that have been sitting around for a couple of hours, as this one had) monocytes tend to aggregate with platelets so they appear more CD61hi. Plus, they express slightly lower levels of CD45 than lymphocytes. In this plot, I think most of the monocytes are in CD61hi region in between the two CD45bright and dim populations. I will admit that the CD61 staining in this sample looks a little different than I'm used to - there are normally more distinct populations of platelet-bound and unbound cells in the monocyte and neutrophil populations (probably just due to the sample sitting around for a long time before staining).

That being said, you are quite right a adding in some additional markers (e.g. CD14) is needed to really tease things out better.

I used CD3, CD19 and CD14 to quickly gate out T cells, B cells and CD14hi monocytes (leaving out most of the CD16hi monocytes). Strangely enough, most of the bright events don't really turn up in the monocyte gate, but instead appear in the B and T cell gates (see first attachment).

140224_staining_issues3.pdf
(316.83 KiB) Downloaded 396 times


Looking at the data more carefully, I think this is because those bright events are actually "spilling over" into several of the Nd and other low mass channels, not just Nd148 (in this panel, that includes Nd142-CD19, Nd145-CD4 and Nd146-CD8) but not into the higher mass channels like Gd160.

140224_staining_issues4.pdf
(356 KiB) Downloaded 363 times
Adeeb Rahman
Icahn School of Medicine at Mount Sinai, NYC
<<

mleipold

Guru

Posts: 5792

Joined: Fri Nov 01, 2013 5:30 pm

Location: Stanford HIMC, CA, USA

Post Fri Mar 21, 2014 4:30 pm

Re: Nd148/Pr141/La139 background issues

Hi Adeeb,

I'm confused why you're now grouping your Granulocytes (large CD45mid grouping) with your Leukocytes (CD45bri) in your staining issues 3 PDF.

See if you can upload the FCS file itself....that way we can play with it ourselves and see whether these are gating artifacts in our hands.


Also, have you titered these antibodies yourself? Even if you buy them from DVS, you should titer them to find the concentration that works best (signal vs background) in *your* hands, in *your* system.


-Mike

Return to CyTOF troubleshooting

Who is online

Users browsing this forum: No registered users and 15 guests