FAQ  •  Register  •  Login

signal shift

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

ianfrank

Contributor

Posts: 29

Joined: Thu Nov 21, 2013 11:59 pm

Location: Tampa, FL

Post Thu Mar 05, 2015 11:03 pm

signal shift

Tube 1 differences.pdf
(156.34 KiB) Downloaded 409 times

Dear All,

I collected an experiment consisting of 13 samples. Sample number one appears to be different than the rest. I've attached some dot plots of tube 1 versus 3 other normal tubes in the collections. Any ideas on what causes the negative events to shift up the axis? This causes a loss in sensitivity, and it has happened before. The speed was the same or slower than the other samples. There was a total of ~370,000 events over 10 minutes. I don't have an autosampler. ....standard cytof one configuration. Also not all channels were effected. 13 of the the 31 channels have this problem, and it's mostly in the 160 to 170 mass range.

Regards,
ian
<<

elinastar

Participant

Posts: 19

Joined: Sun Nov 17, 2013 1:46 pm

Post Fri Mar 06, 2015 5:26 am

Re: signal shift

Hi Ian,

I would suggest to examine the behavior of your positive signal over time. Plot your CD127 vs time and check how strong your positive signal is in tube 1 vs other samples, and does it drift significantly over time of the run . If your positive staining is weaker, your negative population might change. If this is the case, we can share some thoughts on this to help you troubleshoot.


Good luck,
Elina
<<

mleipold

Guru

Posts: 5792

Joined: Fri Nov 01, 2013 5:30 pm

Location: Stanford HIMC, CA, USA

Post Fri Mar 06, 2015 4:46 pm

Re: signal shift

Hi Ian,

I note a few things:

1. If these are 10min injections, then there are some processing issues. 370K (or even 250K) events in 10min is running your samples WAY too fast. You're going to have doublet issues, especially on a CyTOFv1. Your event rate should be closer to 150K, max, in 10min.

2. I notice that your Ir+ events are often less than half of your Ungated events. This argues that you have washing/debris and/or antibody titer problems. If your samples aren't sufficiently washed or your antibody isn't titered well, you'll have higher background in various channels, in the form of speckles/tire tracks in affected channels. Remember, when a cell event is "called" by the algorithm, it grabs the metal signal information from each channel in that event's time window. This means it will grab background from free antibody or debris, even though there's no "true" signal there.

In support of this idea, notice how your CD127 background for CD127-CD3- population decreased in Sample 4 (60% Ir+, 265K Ungated) vs Sample 1 (31% Ir+, 377K Ungated).

Additionally, your Ir signal in Sample 1 is lower than in the other 3 samples: much or most of the Cell Length is driven by Ir staining intensity (since ion cloud size is driven by total metal content, and Ir is usually the highest metal signal in your sample, which helps normalize Cell Length across samples regardless of the antibody staining intensity). This again argues that Sample 1 had a lot of debris, which would have thrown off the Ir/cell ratio and lowered your Ir signal (assuming all these samples were stained with the same Ir dilution stock).


Mike
<<

AdeebR

Grand master

Posts: 169

Joined: Thu Mar 13, 2014 5:58 pm

Location: NYC

Post Fri Mar 06, 2015 10:36 pm

Re: signal shift

I concur with Mike in thinking this is likely because the sample was run at too high an event rate. I've also found that the debris effect that Mike mentions can occur in a well prepared sample simply as a consequence of being run too concentrated. I suppose the fusion of ion clouds at high events rates limits accurate resolution of single cell events and results in a lot of slit doublets, and one of the ways this manifests is that markers start to be inaccurately assigned to cells increasing their apparent background.
Adeeb Rahman
Icahn School of Medicine at Mount Sinai, NYC

Return to CyTOF troubleshooting

Who is online

Users browsing this forum: No registered users and 12 guests

cron