"Optimal" Event Rate

[jimbomahoney]

Hi all,

As I understand it, the maximum event rate on a CyTOF is 1000 e/sec.

Am I right in assuming that higher event rate = higher CV? Is that the only effect (on the data)?

I guess I'm asking if the "optimal" event rate is simply the fastest one can acquire without causing clogs whilst sticking below 1000 e/sec.

Is there a "rule of thumb" for acquisition rates with different samples? e.g. I've heard that "simple" samples, like blood, could be acquired at ~750 e/sec, but samples like dissociated tissue probably need ~250 e/sec, but I'm assuming that this is simply due to the chances of clogs with the tissue.

[mleipold]

Hi James,

There are two issues around event rate.

1. Clogging. Too concentrated of a sample, even polystyrene beads, will increase the likelihood of clogs. For healthy donors, cryoPBMCs can be run a bit faster than whole-blood or bone marrow, which can be run a bit faster than tissue dissociates.

2. Doublets. Too concentrated of a sample, and you'll be throwing out most of your events as doublets. Remember, due to charge repulsion, the ion cloud originating from a cell expands as it travels through the plasma and into the cones. For the HT injector (original narrow-bore original Helios, which I and many others still use), I've heard numbers of around 100x (10um cell -> 1mm ion cloud).

Therefore, it's entirely possible to have cells that are separated while in suspension but whose ion clouds overlap upon charge reuplsion expansion. These events would likely then be read as *one* cell event, not two, and thrown out in various debarcoding and/or Event Length trash removal steps.

As such, one primary way in which Doublet rate is controlled is by having more dilute cell suspensions. We typically run even cryoPBMC at 0.7-0.8M/mL, for a target event rate of 250 events/sec.


Back when we first got the Helios instruments (a bit over 5 yr ago, how time flies), I did do a formal check on recovery, clogging, and doublets. Unfortunately, I can't seem to find that Excel in Cytoforum at the moment. I did discuss some of it here: viewtopic.php?f=1&t=382&p=1336


Mike

[dtelad11]

Adding to Mike's response above: Event rate is "yet another" parameter that you can optimize. Take a typical sample, run it in different event rates, monitor clogging and examine doublet rate.

Suggestions on calculating doublet rate:

- Traditional event length versus DNA, or Gaussian markers.
- Percent of CD3+ CD19+ cells out of [CD3+ or CD19+] cells.
- Split sample into ten tubes. Each tube gets a different CD45 channel (on different mass). Any event that has more than one CD45 is a doublet.

(The last one is courtesy of Adeeb Rahman, and it works like a charm.)

[jimbomahoney]

Thanks both for the explanations.

The clogging I am (sadly!) well aware of, but I'd forgotten the issue of doublets and the ion cloud expansion.

It's interesting (but probably coincidental!) that on a BD Canto II (maximum event rate 10,000), I've heard it recommended that the optimal event rate is around 2,000 - 5,000, and this is essentially the same relative figure (i.e. 20 - 50% of max) as CyTOF.

EDIT - just for fun (I use any excuse to make a plot of something!) I made the following based on a presentation from Fluidigm:



DISCLAIMER - this is obviously just a "back-of-the-envelope" plot and I don't mean to imply it's officially endorsed by Fluidigm, but it's (probably) useful.