FAQ  •  Register  •  Login

future of CyTOF

<<

GregHopkins

Contributor

Posts: 26

Joined: Tue Apr 11, 2017 8:39 pm

Location: 2seventy bio, Cambridge, MA

Post Tue Dec 15, 2020 2:58 pm

future of CyTOF

Hello all,

This post is of a subject that may generate a little bit of controversy. However, science is full of controversy, so I feel as if it is best to address it head on. Its a bit of an "elephant in the room." Moderators, if this is out of scope with what should be posted here, feel free to remove, I won't take it personally :)

I should preface to say I love the CyTOF technology and have poured blood, sweat, and tears into using and furthering the technology - and will continue to do so for as long as my career allows me.

It seems like recently more and more, CyTOF has come under fire from many angles. It has come under fire from the high-parameter flow arena, which touts that flow is more high throughput and commercially available to support high parameter protein detection (but its still fluorescence based, where naturally occurring fluorescence could complicate panel development). It has come under fire from the oligo-conjugated sequencing antibody side, where it is touted that up to 300 targets can be detected at one time, and used in tandem with rna-sequencing to get genomics and proteomics in one package (however intracellular targets are a challenge, but ASAP Seq seems to be an approach addressing this). Multi-omic platforms that integrate all of these approaches into one single stream are being feverishly developed and poised to hit the market maybe next year or so.


These technologies do not have to be diametrically opposed to each other, but sometimes at least from a vendor side, it is conveyed that they rival CyTOF (various conversations I've had with other vendors try to outcompete CyTOF in their pitches). Some pitches have felt almost predatory in trying to get an advantage over CyTOF. I feel that these technologies can work with each other, but sometimes commercial interests make that a bit harder to actually come true. It begs the question, what is going to give?

What I'd like to hear from everyone is:
-What do you think the future of CyTOF is? Is there enough left to develop in order to keep this technology going for the very long term?
-How do you think that CyTOF can be used in conjunction with technologies like spectral flow and CITESeq (rather than opposed to them)?
-When challenged, how do you defend the validity of CyTOF? I'd love to borrow from others wisdom and experience here, as learning to defend it better is definitely a strong interest of mine).

Hopefully I don't ignite too much here, other than productive discussion and a 360 view of where CyTOF is and where it is going. I look forward to hearing all of your experiences and opinions.

Regards,

Greg Hopkins (ghopkins@bluebirdbio.com)
<<

mleipold

Guru

Posts: 5796

Joined: Fri Nov 01, 2013 5:30 pm

Location: Stanford HIMC, CA, USA

Post Tue Dec 15, 2020 11:01 pm

Re: future of CyTOF

Hi Greg,

I think it's an appropriate topic. Hopefully you'll get at least a few responses even as the holiday shutdowns approach (we've technically already begun ours).

I get similar questions from time to time, from scientists and scientific consulting groups. Here are some of my thoughts.

1. Sequencing has a severe problem of "N": until it can match the number of cells that the various cytometries can, there will always be a place for cytometry. By this, I mean that I have a big problem with a lot of scRNAseq papers giving data on only 100s to low 100Ks of cells in their entire study (sometimes a sum of several tens of donors). We *know* from decades of cytometry what the diversity of cells are even in peripheral blood, let alone tissues such as skin. And that's just in healthy people.


2. RNAseq of any sort has the basic, fundamental issue that you're measuring copies of RNA, not copies of protein. While RNA is important (and can do more than simply code for proteins), any biochemistry textbook will demonstrate that it's the proteins that do the majority of the work for a cell. We also know from decades of fundamental biochemistry that protein and RNA levels are *not* necessarily correlated. I think this is something that the RNAseq field is acknowledging more and more, but I've also read tens or hundreds of RNAseq papers and sat through tens or hundreds of lectures where that was ignored or brushed aside.

CITEseq/Abseq is a great step in this direction, to be able to capture both RNA and protein levels. But it still currently limited by the "N" mentioned above.

- addendum: there's also the issue of sequencing dropouts ( https://dx.doi.org/10.1038/s41467-020-14976-9 ).

3. I also think that it needs to be better understood what the assumptions in cytometry vs scRNAseq analysis are. I'm not an RNAseq person so I won't be able to give the fine details (Lars and others, can you chime in?), but your testing statistics (FDR, etc) have some significant differences between scRNAseq (up to 100Ks of cells but 10Ks of reasonably not-pre-selected probes) and cytometry (100Ks to 10Ms or more of total cell events, but 10s of probes largely pre-selected by decades of immunology and biochemistry knowledge).

Things like Florian Mair's "Immunome" approach (https://doi.org/10.1101/700534 ; http://dx.doi.org/10.1016/j.celrep.2020.03.063 ) balance this a bit more.


4. Fluorescent flow (spectral or traditional) has a strong throughput advantage to CyTOF. Also, its ability to stain and sort live cells for further analysis/culture is also an advantage over CyTOF. However, CyTOF still currently has the advantage of number of possible probes. And as you mention, there are also still autofluorescence and spillover in flow (spillover is still generally less in CyTOF).


I still largely see these as platforms helping to address different levels of focus: for example, CyTOF (or CITEseq) allow you to cast your net broadly, especially in cases where you're not already sure what to look for. Once you've convinced yourself (ideally through multiple independent platforms!) that you have something of interest, fluorescent flow is far and away the best way to either crank through thousands of study samples, and/or develop into a clinical diagnostic assay.


Mike
<<

dtelad11

Master

Posts: 129

Joined: Mon Oct 31, 2016 6:26 pm

Post Wed Dec 23, 2020 3:02 pm

Re: future of CyTOF

I think that Mike summarized my opinion well:

I still largely see these as platforms helping to address different levels of focus: for example, CyTOF (or CITEseq) allow you to cast your net broadly, especially in cases where you're not already sure what to look for. Once you've convinced yourself (ideally through multiple independent platforms!) that you have something of interest, fluorescent flow is far and away the best way to either crank through thousands of study samples, and/or develop into a clinical diagnostic assay.


I view the CyTOF as a powerful discovery engine. Given the ease of developing a new panel, the high marker count, and the availability of barcoding, I believe it's currently the best tool for assaying a reasonable number of events from a limited number of samples (10s to low 100s). Then you can proceed with a tailored flow or spectral cytometry panel.

With regards to CITEseq, Mike outlined a few technical challenges with the technology. Setting them aside, I agree that there is overlap with CyTOF with regards to discovery capabilities. They each have their nuances and I think a skilled researcher will pick the right technology for her or his experiment.

I was curious about Greg's last question,

-When challenged, how do you defend the validity of CyTOF? I'd love to borrow from others wisdom and experience here, as learning to defend it better is definitely a strong interest of mine).


Why do you feel the need the defend the validity of the CyTOF? In my opinion, that's Fluidigm's job. Mass cytometry is a piece of technology, like any other. It has its strengths and weaknesses. It fits some experiments well, and others -- not so much.
<<

GregHopkins

Contributor

Posts: 26

Joined: Tue Apr 11, 2017 8:39 pm

Location: 2seventy bio, Cambridge, MA

Post Mon Feb 08, 2021 8:27 pm

Re: future of CyTOF

Mike and El-Ad,

Thank you as always for your insights and views on the state of all of the -omics technologies. You both always have very insightful opinions to offer. I'm becoming more and more aware of the limits and benefits of many of the various -omics technologies, and as of now, they all seem to have a fairly clear place based on the researcher's needs.

However, all technologies evolve, as they should. One thing I always wonder is how mass cytometry (which is admittedly close to my heart) will evolve? What room is there to continually innovate or expand the capabilities in a marketable way? I personally don't know but am very interested to see where it goes and hopefully participate in it. I have heard an anecdote once before that was to the effect of it not being about pushing past 40 or 50 parameters, but what you can do with those 40 to 50 parameters. I think that's what excites me, utilizing those capabilities.

Mass cytometry may not be THE most comprehensive proteomic technology out there, but I do feel it offers some things that other proteomic technologies don't. As I expressed, they shouldn't compete with one another - its much more productive if one fills in the gaps that exist in others to create a comprehensive package. The genomic/proteomic-in-one developments that are in the field now are really exciting and definitely have their place - they produce good data just like any other platform when used wisely.

Mass cytometry has its nay-sayers and its proponents, and some that are in between. There are lots of reasons to be a proponent of mass cytometry, and I guess you could say I'm invested in being a steadfast proponent.

Looking forward to keeping these types of discussions alive, in whatever form they may come :)

Return to CyTOF general discussion

Who is online

Users browsing this forum: No registered users and 15 guests