Page 1 of 1

Doublets vs. G2 cells

PostPosted: Mon May 11, 2020 3:33 pm
by gerokn
Hi all,

I have a rather simple question: How do you actually gate single cells based on the Ir signal? Usually, there is a big population representing the singlets and a second, smaller population supposedly representing the doublets. But is this "doublet" population not consisting of both doublets of G1 cells as well as single G2 cells? If one prepares a well dispersed cell suspension with a relatively low concentration, would one harm the dataset more by removing G2 cells than actually correcting it?

Just to hear your thoughts and how you handle this. I was taught to gate on that "G1 population" but I'm not sure that's a good procedure.

Best,
Gero

Re: Doublets vs. G2 cells

PostPosted: Mon May 11, 2020 5:51 pm
by dtelad11

Re: Doublets vs. G2 cells

PostPosted: Mon May 11, 2020 6:00 pm
by mleipold
Hi Gero,

In addition to El-ad's suggestion, here are a few other papers that might be relevant to you:

Tsai et al: https://www.ncbi.nlm.nih.gov/pubmed/32161403
Behbehani: https://www.ncbi.nlm.nih.gov/pubmed/31077107
Stern et al: https://www.ncbi.nlm.nih.gov/pubmed/27768827
Good et al: https://www.ncbi.nlm.nih.gov/pubmed/30742126


Mike

Re: Doublets vs. G2 cells

PostPosted: Mon May 11, 2020 6:26 pm
by dtelad11
Mike, next time we meet in a conference, I want to play a game where someone asks a CyTOF question and each contestant writes a list of proposed manuscripts. Winner is the one that offers the most reading options ;)

Re: Doublets vs. G2 cells

PostPosted: Wed May 13, 2020 9:35 am
by idurei
Hi Gero,

It's nice to see that more people also recognize this problem.

In Fig 2 of our paper, suggested by El-ad (https://onlinelibrary.wiley.com/doi/ful ... to.a.23960) , we show how G1+G1 aggregates are completely overlapping with M phase cells in all the conventional 2D gating plots used for single-cell gating of mass cytometry data. Thus, you will exclude cells late in the cycle when excluding G1+G1 aggregates in ELvsIr, Ir191vsIr193 or ELvsCisplatin. Allthough the new Gaussian parameters in CyTOF software v 7 is quite good at excluding ion cloud fusions (coincidence), true cell aggregates from the sample (stuck together) will still be present Ion clouds from an aggregate still have a Gaussian shaped pulse.

We have used cell cycle markers (anti-CDT1, IdU, anti-Geminin and anti-PhosphoHistoneH3) to do single cell gating on each cell cycle phase individually (see Fig1 E). When doing single cell gating this way there are no significant differences between cell cycle analysis by flow cytometry and mass cytometry of identical samples. Even if you are not interested in cell cycle per se, it would be useful to include these marker just to do be able to perform proper single cell gating without exclusion of single cells in late S/G2/M. Ir-resolution is often poor (but still proportianl to DNA content), so we have used Pd-barcode intensity (relative to cell volume) vs Cisplatin (relative to cell surface area) to have a uniform single-cell gating strategy for all the cell cycle phases, no matter the Ir-resolution.

Another problem is that some cell types (e.g. monocytes) bind more Ir than others (e.g. lymphocytes) due to the chromatin structure. So strict gating on Ir would exclude monocytes in PBMC analysis by mass cytometry (see viewtopic.php?f=3&t=392&hilit=monocytes+Iridium).

Idun

Re: Doublets vs. G2 cells

PostPosted: Fri May 15, 2020 6:24 am
by gerokn
Thanks very much to both of you! That helps a lot.

Have a great weekend,
Gero